Paper Digest: EMNLP 2020 (Main Track) Highlights
The Conference on Empirical Methods in Natural Language Processing (EMNLP) is one of the top natural language processing conferences in the world. In 2020, it is to be held online due to covid-19 pandemic.
An innovation for EMNLP 2020 is a new acceptance category, which will allow for more high quality papers (short and long) to be accepted than usual. EMNLP 2020 is creating a new sister publication, Findings of ACL: EMNLP 2020 (hereafter Findings), which will serve as an online companion publication for papers that are not accepted for publication in the main conference, but nonetheless have been assessed by the programme committee as solid work with sufficient substance, quality and novelty to warrant publication. All these Finding track papers are put in a seperate page: Findings Track Paper Highlights.
Readers can choose to read all EMNLP-2020 papers including both main track and findings track on our console, which allows users to filter out papers using keywords and find related papers and patents.
To help the community quickly catch up on the work presented in this conference, Paper Digest Team processed all accepted papers, and generated one highlight sentence (typically the main topic) for each paper. Readers are encouraged to read these machine generated highlights / summaries to quickly get the main idea of each paper.
If you do not want to miss any interesting academic paper, you are welcome to sign up our free daily paper digest service to get updates on new papers published in your area every day. You are also welcome to follow us on Twitter and Linkedin to get updated with new conference digests.
Paper Digest Team
team@paperdigest.org
TABLE 1: Paper Digest: EMNLP 2020 (Main Track) Highlights
Paper | Author(s) | Code | |
---|---|---|---|
1 | Detecting Attackable Sentences In Arguments Highlight: We present a first large-scale analysis of sentence attackability in online arguments. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Yohan Jo; Seojin Bang; Emaad Manzoor; Eduard Hovy; Chris Reed; | |
2 | Extracting Implicitly Asserted Propositions In Argumentation Highlight: In this paper, we examine a wide range of computational methods for extracting propositions that are implicitly asserted in questions, reported speech, and imperatives in argumentation. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Yohan Jo; Jacky Visser; Chris Reed; Eduard Hovy; | |
3 | Quantitative Argument Summarization And Beyond: Cross-domain Key Point Analysis Highlight: The current work advances key point analysis in two important respects: first, we develop a method for automatic extraction of key points, which enables fully automatic analysis, and is shown to achieve performance comparable to a human expert. Second, we demonstrate that the applicability of key point analysis goes well beyond argumentation data. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Roy Bar-Haim; Yoav Kantor; Lilach Eden; Roni Friedman; Dan Lahav; Noam Slonim; | |
4 | Unsupervised Stance Detection For Arguments From Consequences Highlight: In this paper, we propose an unsupervised method to detect the stance of argumentative claims with respect to a topic. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Jonathan Kobbe; Ioana Hulpuș; Heiner Stuckenschmidt; | |
5 | BLEU Might Be Guilty But References Are Not Innocent Highlight: We study different methods to collect references and compare their value in automated evaluation by reporting correlation with human evaluation for a variety of systems and metrics. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Markus Freitag; David Grangier; Isaac Caswell; | |
6 | Statistical Power And Translationese In Machine Translation Evaluation Highlight: The term translationese has been used to describe features of translated text, and in this paper, we provide detailed analysis of potential adverse effects of translationese on machine translation evaluation. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Yvette Graham; Barry Haddow; Philipp Koehn; | |
7 | Simulated Multiple Reference Training Improves Low-resource Machine Translation Highlight: We introduce Simulated Multiple Reference Training (SMRT), a novel MT training method that approximates the full space of possible translations by sampling a paraphrase of the reference sentence from a paraphraser and training the MT model to predict the paraphraser’s distribution over possible tokens. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Huda Khayrallah; Brian Thompson; Matt Post; Philipp Koehn; | |
8 | Automatic Machine Translation Evaluation In Many Languages Via Zero-Shot Paraphrasing Highlight: We propose training the paraphraser as a multilingual NMT system, treating paraphrasing as a zero-shot translation task (e.g., Czech to Czech). Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Brian Thompson; Matt Post; | |
9 | PRover: Proof Generation For Interpretable Reasoning Over Rules Highlight: In our work, we take a step closer to emulating formal theorem provers, by proposing PRover, an interpretable transformer-based model that jointly answers binary questions over rule-bases and generates the corresponding proofs. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Swarnadeep Saha; Sayan Ghosh; Shashank Srivastava; Mohit Bansal; | |
10 | Learning To Explain: Datasets And Models For Identifying Valid Reasoning Chains In Multihop Question-Answering Highlight: To address this, we introduce three explanation datasets in which explanations formed from corpus facts are annotated. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Harsh Jhamtani; Peter Clark; | |
11 | Self-Supervised Knowledge Triplet Learning For Zero-Shot Question Answering Highlight: This work proposes Knowledge Triplet Learning (KTL), a self-supervised task over knowledge graphs. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Pratyay Banerjee; Chitta Baral; | |
12 | More Bang For Your Buck: Natural Perturbation For Robust Question Answering Highlight: As an alternative to the traditional approach of creating new instances by repeating the process of creating one instance, we propose doing so by first collecting a set of seed examples and then applying human-driven natural perturbations (as opposed to rule-based machine perturbations), which often change the gold label as well. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Daniel Khashabi; Tushar Khot; Ashish Sabharwal; | |
13 | A Matter Of Framing: The Impact Of Linguistic Formalism On Probing Results Highlight: To investigate, we conduct an in-depth cross-formalism layer probing study in role semantics. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Ilia Kuznetsov; Iryna Gurevych; | |
14 | Information-Theoretic Probing With Minimum Description Length Highlight: Instead, we propose an alternative to the standard probes, information-theoretic probing with minimum description length (MDL). Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Elena Voita; Ivan Titov; | |
15 | Intrinsic Probing Through Dimension Selection Highlight: To enable intrinsic probing, we propose a novel framework based on a decomposable multivariate Gaussian probe that allows us to determine whether the linguistic information in word embeddings is dispersed or focal. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Lucas Torroba Hennigen; Adina Williams; Ryan Cotterell; | |
16 | Learning Which Features Matter: RoBERTa Acquires A Preference For Linguistic Generalizations (Eventually) Highlight: With this goal in mind, we introduce a new English-language diagnostic set called MSGS (the Mixed Signals Generalization Set), which consists of 20 ambiguous binary classification tasks that we use to test whether a pretrained model prefers linguistic or surface generalizations during finetuning. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Alex Warstadt; Yian Zhang; Xiaocheng Li; Haokun Liu; Samuel R. Bowman; | |
17 | Repulsive Attention: Rethinking Multi-head Attention As Bayesian Inference Highlight: In this paper, for the first time, we provide a novel understanding of multi-head attention from a Bayesian perspective. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Bang An; Jie Lyu; Zhenyi Wang; Chunyuan Li; Changwei Hu; Fei Tan; Ruiyi Zhang; Yifan Hu; Changyou Chen; | |
18 | KERMIT: Complementing Transformer Architectures With Encoders Of Explicit Syntactic Interpretations Highlight: In this paper, we propose KERMIT (Kernel-inspired Encoder with Recursive Mechanism for Interpretable Trees) to embed symbolic syntactic parse trees into artificial neural networks and to visualize how syntax is used in inference. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Fabio Massimo Zanzotto; Andrea Santilli; Leonardo Ranaldi; Dario Onorati; Pierfrancesco Tommasino; Francesca Fallucchi; | |
19 | ETC: Encoding Long And Structured Inputs In Transformers Highlight: In this paper, we present a new Transformer architecture, Extended Transformer Construction (ETC), that addresses two key challenges of standard Transformer architectures, namely scaling input length and encoding structured inputs. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Joshua Ainslie; Santiago Ontanon; Chris Alberti; Vaclav Cvicek; Zachary Fisher; Philip Pham; Anirudh Ravula; Sumit Sanghai; Qifan Wang; Li Yang; | |
20 | Pre-Training Transformers As Energy-Based Cloze Models Highlight: We introduce Electric, an energy-based cloze model for representation learning over text. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Kevin Clark; Minh-Thang Luong; Quoc Le; Christopher D. Manning; | |
21 | Calibration Of Pre-trained Transformers Highlight: We focus on BERT and RoBERTa in this work, and analyze their calibration across three tasks: natural language inference, paraphrase detection, and commonsense reasoning. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Shrey Desai; Greg Durrett; | |
22 | Near-imperceptible Neural Linguistic Steganography Via Self-Adjusting Arithmetic Coding Highlight: In this study, we present a new linguistic steganography method which encodes secret messages using self-adjusting arithmetic coding based on a neural language model. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Jiaming Shen; Heng Ji; Jiawei Han; | |
23 | Multi-Dimensional Gender Bias Classification Highlight: In this work, we propose a novel, general framework that decomposes gender bias in text along several pragmatic and semantic dimensions: bias from the gender of the person being spoken about, bias from the gender of the person being spoken to, and bias from the gender of the speaker. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Emily Dinan; Angela Fan; Ledell Wu; Jason Weston; Douwe Kiela; Adina Williams; | |
24 | FIND: Human-in-the-Loop Debugging Deep Text Classifiers Highlight: In this paper, we propose FIND – a framework which enables humans to debug deep learning text classifiers by disabling irrelevant hidden features. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Piyawat Lertvittayakumjorn; Lucia Specia; Francesca Toni; | |
25 | Conversational Document Prediction To Assist Customer Care Agents Highlight: We study the task of predicting the documents that customer care agents can use to facilitate users’ needs. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Jatin Ganhotra; Haggai Roitman; Doron Cohen; Nathaniel Mills; Chulaka Gunasekara; Yosi Mass; Sachindra Joshi; Luis Lastras; David Konopnicki; | |
26 | Incremental Processing In The Age Of Non-Incremental Encoders: An Empirical Assessment Of Bidirectional Models For Incremental NLU Highlight: We investigate how they behave under incremental interfaces, when partial output must be provided based on partial input seen up to a certain time step, which may happen in interactive systems. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Brielen Madureira; David Schlangen; | |
27 | Augmented Natural Language For Generative Sequence Labeling Highlight: We propose a generative framework for joint sequence labeling and sentence-level classification. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Ben Athiwaratkun; Cicero Nogueira dos Santos; Jason Krone; Bing Xiang; | |
28 | Dialogue Response Ranking Training With Large-Scale Human Feedback Data Highlight: We leverage social media feedback data (number of replies and upvotes) to build a large-scale training dataset for feedback prediction. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Xiang Gao; Yizhe Zhang; Michel Galley; Chris Brockett; Bill Dolan; | |
29 | Semantic Evaluation For Text-to-SQL With Distilled Test Suites Highlight: We propose test suite accuracy to approximate semantic accuracy for Text-to-SQL models. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Ruiqi Zhong; Tao Yu; Dan Klein; | |
30 | Cross-Thought For Sentence Encoder Pre-training Highlight: In this paper, we propose Cross-Thought, a novel approach to pre-training sequence encoder, which is instrumental in building reusable sequence embeddings for large-scale NLP tasks such as question answering. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Shuohang Wang; Yuwei Fang; Siqi Sun; Zhe Gan; Yu Cheng; Jingjing Liu; Jing Jiang; | |
31 | AutoQA: From Databases To QA Semantic Parsers With Only Synthetic Training Data Highlight: We propose AutoQA, a methodology and toolkit to generate semantic parsers that answer questions on databases, with no manual effort. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Silei Xu; Sina Semnani; Giovanni Campagna; Monica Lam; | |
32 | A Spectral Method For Unsupervised Multi-Document Summarization Highlight: In this paper, we propose a spectral-based hypothesis, which states that the goodness of summary candidate is closely linked to its so-called spectral impact. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Kexiang Wang; Baobao Chang; Zhifang Sui; | |
33 | What Have We Achieved On Text Summarization? Highlight: Aiming to gain more understanding of summarization systems with respect to their strengths and limits on a fine-grained syntactic and semantic level, we consult the Multidimensional Quality Metric (MQM) and quantify 8 major sources of errors on 10 representative summarization models manually. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Dandan Huang; Leyang Cui; Sen Yang; Guangsheng Bao; Kun Wang; Jun Xie; Yue Zhang; | |
34 | Q-learning With Language Model For Edit-based Unsupervised Summarization Highlight: In this paper, we propose a new approach based on Q-learning with an edit-based summarization. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Ryosuke Kohita; Akifumi Wachi; Yang Zhao; Ryuki Tachibana; | |
35 | Friendly Topic Assistant For Transformer Based Abstractive Summarization Highlight: To this end, we rearrange and explore the semantics learned by a topic model, and then propose a topic assistant (TA) including three modules. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Zhengjue Wang; Zhibin Duan; Hao Zhang; Chaojie Wang; Long Tian; Bo Chen; Mingyuan Zhou; | |
36 | Contrastive Distillation On Intermediate Representations For Language Model Compression Highlight: To achieve better distillation efficacy, we propose Contrastive Distillation on Intermediate Representations (CoDIR), a principled knowledge distillation framework where the student is trained to distill knowledge through intermediate layers of the teacher via a contrastive objective. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Siqi Sun; Zhe Gan; Yuwei Fang; Yu Cheng; Shuohang Wang; Jingjing Liu; | |
37 | TernaryBERT: Distillation-aware Ultra-low Bit BERT Highlight: In this work, we propose TernaryBERT, which ternarizes the weights in a fine-tuned BERT model. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Wei Zhang; Lu Hou; Yichun Yin; Lifeng Shang; Xiao Chen; Xin Jiang; Qun Liu; | |
38 | Self-Supervised Meta-Learning For Few-Shot Natural Language Classification Tasks Highlight: This paper proposes a self-supervised approach to generate a large, rich, meta-learning task distribution from unlabeled text. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Trapit Bansal; Rishikesh Jha; Tsendsuren Munkhdalai; Andrew McCallum; | |
39 | Efficient Meta Lifelong-Learning With Limited Memory Highlight: In this paper, we identify three common principles of lifelong learning methods and propose an efficient meta-lifelong framework that combines them in a synergistic fashion. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Zirui Wang; Sanket Vaibhav Mehta; Barnabas Poczos; Jaime Carbonell; | |
40 | Don’t Use English Dev: On The Zero-Shot Cross-Lingual Evaluation Of Contextual Embeddings Highlight: We show that the standard practice of using English dev accuracy for model selection in the zero-shot setting makes it difficult to obtain reproducible results on the MLDoc and XNLI tasks. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Phillip Keung; Yichao Lu; Julian Salazar; Vikas Bhardwaj; | |
41 | A Supervised Word Alignment Method Based On Cross-Language Span Prediction Using Multilingual BERT Highlight: We present a novel supervised word alignment method based on cross-language span prediction. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Masaaki Nagata; Katsuki Chousa; Masaaki Nishino; | |
42 | Accurate Word Alignment Induction From Neural Machine Translation Highlight: In this paper, we show that attention weights do capture accurate word alignments and propose two novel word alignment induction methods Shift-Att and Shift-AET. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Yun Chen; Yang Liu; Guanhua Chen; Xin Jiang; Qun Liu; | |
43 | ChrEn: Cherokee-English Machine Translation For Endangered Language Revitalization Highlight: To help save this endangered language, we introduce ChrEn, a Cherokee-English parallel dataset, to facilitate machine translation research between Cherokee and English. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Shiyue Zhang; Benjamin Frey; Mohit Bansal; | |
44 | Unsupervised Discovery Of Implicit Gender Bias Highlight: We take an unsupervised approach to identifying gender bias against women at a comment level and present a model that can surface text likely to contain bias. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Anjalie Field; Yulia Tsvetkov; | |
45 | Condolence And Empathy In Online Communities Highlight: Here, we develop computational tools to create a massive dataset of 11.4M expressions of distress and 2.8M corresponding offerings of condolence in order to examine the dynamics of condolence online. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Naitian Zhou; David Jurgens; | |
46 | An Embedding Model For Estimating Legislative Preferences From The Frequency And Sentiment Of Tweets Highlight: In this paper we introduce a method of measuring more specific legislator attitudes using an alternative expression of preferences: tweeting. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Gregory Spell; Brian Guay; Sunshine Hillygus; Lawrence Carin; | |
47 | Measuring Information Propagation In Literary Social Networks Highlight: We describe a new pipeline for measuring information propagation in this domain and publish a new dataset for speaker attribution, enabling the evaluation of an important component of this pipeline on a wider range of literary texts than previously studied. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Matthew Sims; David Bamman; | |
48 | Social Chemistry 101: Learning To Reason About Social And Moral Norms Highlight: We present SOCIAL CHEMISTRY, a new conceptual formalism to study people’s everyday social norms and moral judgments over a rich spectrum of real life situations described in natural language. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Maxwell Forbes; Jena D. Hwang; Vered Shwartz; Maarten Sap; Yejin Choi; | |
49 | Event Extraction By Answering (Almost) Natural Questions Highlight: To avoid this issue, we introduce a new paradigm for event extraction by formulating it as a question answering (QA) task that extracts the event arguments in an end-to-end manner. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Xinya Du; Claire Cardie; | |
50 | Connecting The Dots: Event Graph Schema Induction With Path Language Modeling Highlight: We propose a new Event Graph Schema, where two event types are connected through multiple paths involving entities that fill important roles in a coherent story. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Manling Li; Qi Zeng; Ying Lin; Kyunghyun Cho; Heng Ji; Jonathan May; Nathanael Chambers; Clare Voss; | |
51 | Joint Constrained Learning For Event-Event Relation Extraction Highlight: Due to the lack of jointly labeled data for these relational phenomena and the restriction on the structures they articulate, we propose a joint constrained learning framework for modeling event-event relations. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Haoyu Wang; Muhao Chen; Hongming Zhang; Dan Roth; | |
52 | Incremental Event Detection Via Knowledge Consolidation Networks Highlight: In this paper, we propose a Knowledge Consolidation Network (KCN) to address the above issues. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Pengfei Cao; Yubo Chen; Jun Zhao; Taifeng Wang; | |
53 | Semi-supervised New Event Type Induction And Event Detection Highlight: In this paper, we work on a new task of semi-supervised event type induction, aiming to automatically discover a set of unseen types from a given corpus by leveraging annotations available for a few seen types. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Lifu Huang; Heng Ji; | |
54 | Language Generation With Multi-Hop Reasoning On Commonsense Knowledge Graph Highlight: In this paper, we propose Generation with Multi-Hop Reasoning Flow (GRF) that enables pre-trained models with dynamic multi-hop reasoning on multi-relational paths extracted from the external commonsense knowledge graph. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Haozhe Ji; Pei Ke; Shaohan Huang; Furu Wei; Xiaoyan Zhu; Minlie Huang; | |
55 | Reformulating Unsupervised Style Transfer As Paraphrase Generation Highlight: In this paper, we reformulate unsupervised style transfer as a paraphrase generation problem, and present a simple methodology based on fine-tuning pretrained language models on automatically generated paraphrase data. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Kalpesh Krishna; John Wieting; Mohit Iyyer; | |
56 | De-Biased Court’s View Generation With Causality Highlight: In this paper, we propose a novel Attentional and Counterfactual based Natural Language Generation (AC-NLG) method, consisting of an attentional encoder and a pair of innovative counterfactual decoders. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Yiquan Wu; Kun Kuang; Yating Zhang; Xiaozhong Liu; Changlong Sun; Jun Xiao; Yueting Zhuang; Luo Si; Fei Wu; | |
57 | PAIR: Planning And Iterative Refinement In Pre-trained Transformers For Long Text Generation Highlight: In this work, we present a novel content-controlled text generation framework, PAIR, with planning and iterative refinement, which is built upon a large model, BART. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Xinyu Hua; Lu Wang; | |
58 | Back To The Future: Unsupervised Backprop-based Decoding For Counterfactual And Abductive Commonsense Reasoning Highlight: In this paper, we propose DeLorean, a new unsupervised decoding algorithm that can flexibly incorporate both the past and future contexts using only off-the-shelf, left-to-right language models and no supervision. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Lianhui Qin; Vered Shwartz; Peter West; Chandra Bhagavatula; Jena D. Hwang; Ronan Le Bras; Antoine Bosselut; Yejin Choi; | |
59 | Where Are You? Localization From Embodied Dialog Highlight: In this paper, we focus on the LED task – providing a strong baseline model with detailed ablations characterizing both dataset biases and the importance of various modeling choices. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Meera Hahn; Jacob Krantz; Dhruv Batra; Devi Parikh; James Rehg; Stefan Lee; Peter Anderson; | |
60 | Learning To Represent Image And Text With Denotation Graph Highlight: In this paper, we propose learning representations from a set of implied, visually grounded expressions between image and text, automatically mined from those datasets. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Bowen Zhang; Hexiang Hu; Vihan Jain; Eugene Ie; Fei Sha; | code |
61 | Video2Commonsense: Generating Commonsense Descriptions To Enrich Video Captioning Highlight: We present the first work on generating \textit{commonsense} captions directly from videos, to describe latent aspects such as intentions, effects, and attributes. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Zhiyuan Fang; Tejas Gokhale; Pratyay Banerjee; Chitta Baral; Yezhou Yang; | |
62 | Does My Multimodal Model Learn Cross-modal Interactions? It’s Harder To Tell Than You Might Think! Highlight: We propose a new diagnostic tool, empirical multimodally-additive function projection (EMAP), for isolating whether or not cross-modal interactions improve performance for a given model on a given task. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Jack Hessel; Lillian Lee; | |
63 | MUTANT: A Training Paradigm For Out-of-Distribution Generalization In Visual Question Answering Highlight: In this paper, we present \textit{MUTANT}, a training paradigm that exposes the model to perceptually similar, yet semantically distinct \textit{mutations} of the input, to improve OOD generalization, such as the VQA-CP challenge. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Tejas Gokhale; Pratyay Banerjee; Chitta Baral; Yezhou Yang; | |
64 | Mitigating Gender Bias For Neural Dialogue Generation With Adversarial Learning Highlight: In this paper, we propose a novel adversarial learning framework Debiased-Chat to train dialogue models free from gender bias while keeping their performance. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Haochen Liu; Wentao Wang; Yiqi Wang; Hui Liu; Zitao Liu; Jiliang Tang; | |
65 | Will I Sound Like Me? Improving Persona Consistency In Dialogues Through Pragmatic Self-Consciousness Highlight: We explore the task of improving persona consistency of dialogue agents. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Hyunwoo Kim; Byeongchang Kim; Gunhee Kim; | |
66 | TOD-BERT: Pre-trained Natural Language Understanding For Task-Oriented Dialogue Highlight: In this work, we unify nine human-human and multi-turn task-oriented dialogue datasets for language modeling. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Chien-Sheng Wu; Steven C.H. Hoi; Richard Socher; Caiming Xiong; | |
67 | RiSAWOZ: A Large-Scale Multi-Domain Wizard-of-Oz Dataset With Rich Semantic Annotations For Task-Oriented Dialogue Modeling Highlight: In order to alleviate the shortage of multi-domain data and to capture discourse phenomena for task-oriented dialogue modeling, we propose RiSAWOZ, a large-scale multi-domain Chinese Wizard-of-Oz dataset with Rich Semantic Annotations. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Jun Quan; Shian Zhang; Qian Cao; Zizhong Li; Deyi Xiong; | |
68 | Filtering Noisy Dialogue Corpora By Connectivity And Content Relatedness Highlight: In this paper, we propose a method for scoring the quality of utterance pairs in terms of their connectivity and relatedness. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Reina Akama; Sho Yokoi; Jun Suzuki; Kentaro Inui; | |
69 | Latent Geographical Factors For Analyzing The Evolution Of Dialects In Contact Highlight: In this paper, we propose a probabilistic generative model that represents latent factors as geographical distributions. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Yugo Murawaki; | |
70 | Predicting Reference: What Do Language Models Learn About Discourse Models? Highlight: We address this question by drawing on a rich psycholinguistic literature that has established how different contexts affect referential biases concerning who is likely to be referred to next. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Shiva Upadhye; Leon Bergen; Andrew Kehler; | |
71 | Word Class Flexibility: A Deep Contextualized Approach Highlight: We propose a principled methodology to explore regularity in word class flexibility. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Bai Li; Guillaume Thomas; Yang Xu; Frank Rudzicz; | |
72 | Shallow-to-Deep Training For Neural Machine Translation Highlight: In this paper, we investigate the behavior of a well-tuned deep Transformer system. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Bei Li; Ziyang Wang; Hui Liu; Yufan Jiang; Quan Du; Tong Xiao; Huizhen Wang; Jingbo Zhu; | code |
73 | Iterative Refinement In The Continuous Space For Non-Autoregressive Neural Machine Translation Highlight: We propose an efficient inference procedure for non-autoregressive machine translation that iteratively refines translation purely in the continuous space. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Jason Lee; Raphael Shu; Kyunghyun Cho; | |
74 | Why Skip If You Can Combine: A Simple Knowledge Distillation Technique For Intermediate Layers Highlight: In this paper, we target low-resource settings and evaluate our translation engines for Portuguese?English, Turkish?English, and English?German directions. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Yimeng Wu; Peyman Passban; Mehdi Rezagholizadeh; Qun Liu; | |
75 | Multi-task Learning For Multilingual Neural Machine Translation Highlight: In this work, we propose a multi-task learning (MTL) framework that jointly trains the model with the translation task on bitext data and two denoising tasks on the monolingual data. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Yiren Wang; ChengXiang Zhai; Hany Hassan; | |
76 | Token-level Adaptive Training For Neural Machine Translation Highlight: In this paper, we explored target token-level adaptive objectives based on token frequencies to assign appropriate weights for each target token during training. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Shuhao Gu; Jinchao Zhang; Fandong Meng; Yang Feng; Wanying Xie; Jie Zhou; Dong Yu; | |
77 | Multi-Unit Transformers For Neural Machine Translation Highlight: In this paper, we propose the Multi-Unit Transformer (MUTE) , which aim to promote the expressiveness of the Transformer by introducing diverse and complementary units. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Jianhao Yan; Fandong Meng; Jie Zhou; | |
78 | On The Sparsity Of Neural Machine Translation Models Highlight: In response to this problem, we empirically investigate whether the redundant parameters can be reused to achieve better performance. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Yong Wang; Longyue Wang; Victor Li; Zhaopeng Tu; | |
79 | Incorporating A Local Translation Mechanism Into Non-autoregressive Translation Highlight: In this work, we introduce a novel local autoregressive translation (LAT) mechanism into non-autoregressive translation (NAT) models so as to capture local dependencies among target outputs. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Xiang Kong; Zhisong Zhang; Eduard Hovy; | |
80 | Self-Paced Learning For Neural Machine Translation Highlight: We ameliorate this procedure with a more flexible manner by proposing self-paced learning, where NMT model is allowed to 1) automatically quantify the learning confidence over training examples; and 2) flexibly govern its learning via regulating the loss in each iteration step. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Yu Wan; Baosong Yang; Derek F. Wong; Yikai Zhou; Lidia S. Chao; Haibo Zhang; Boxing Chen; | |
81 | Long-Short Term Masking Transformer: A Simple But Effective Baseline For Document-level Neural Machine Translation Highlight: In this paper, we research extensively the pros and cons of the standard transformer in document-level translation, and find that the auto-regressive property can simultaneously bring both the advantage of the consistency and the disadvantage of error accumulation. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Pei Zhang; Boxing Chen; Niyu Ge; Kai Fan; | |
82 | Generating Diverse Translation From Model Distribution With Dropout Highlight: In this paper, we propose to generate diverse translations by deriving a large number of possible models with Bayesian modelling and sampling models from them for inference. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Xuanfu Wu; Yang Feng; Chenze Shao; | |
83 | Non-Autoregressive Machine Translation With Latent Alignments Highlight: This paper presents two strong methods, CTC and Imputer, for non-autoregressive machine translation that model latent alignments with dynamic programming. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Chitwan Saharia; William Chan; Saurabh Saxena; Mohammad Norouzi; | |
84 | Look At The First Sentence: Position Bias In Question Answering Highlight: In this study, we hypothesize that when the distribution of the answer positions is highly skewed in the training set (e.g., answers lie only in the k-th sentence of each passage), QA models predicting answers as positions can learn spurious positional cues and fail to give answers in different positions. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Miyoung Ko; Jinhyuk Lee; Hyunjae Kim; Gangwoo Kim; Jaewoo Kang; | |
85 | ProtoQA: A Question Answering Dataset For Prototypical Common-Sense Reasoning Highlight: This paper introduces a new question answering dataset for training and evaluating common sense reasoning capabilities of artificial intelligence systems in such prototypical situations. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Michael Boratko; Xiang Li; Tim O’Gorman; Rajarshi Das; Dan Le; Andrew McCallum; | |
86 | IIRC: A Dataset Of Incomplete Information Reading Comprehension Questions Highlight: To fill this gap, we present a dataset, IIRC, with more than 13K questions over paragraphs from English Wikipedia that provide only partial information to answer them, with the missing information occurring in one or more linked documents. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
James Ferguson; Matt Gardner; Hannaneh Hajishirzi; Tushar Khot; Pradeep Dasigi; | code |
87 | Unsupervised Adaptation Of Question Answering Systems Via Generative Self-training Highlight: In this paper we investigate the iterative generation of synthetic QA pairs as a way to realize unsupervised self adaptation. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Steven Rennie; Etienne Marcheret; Neil Mallinar; David Nahamoo; Vaibhava Goel; | |
88 | TORQUE: A Reading Comprehension Dataset Of Temporal Ordering Questions Highlight: We introduce TORQUE, a new English reading comprehension benchmark built on 3.2k news snippets with 21k human-generated questions querying temporal relationships. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Qiang Ning; Hao Wu; Rujun Han; Nanyun Peng; Matt Gardner; Dan Roth; | |
89 | ToTTo: A Controlled Table-To-Text Generation Dataset Highlight: We present ToTTo, an open-domain English table-to-text dataset with over 120,000 training examples that proposes a controlled generation task: given a Wikipedia table and a set of highlighted table cells, produce a one-sentence description. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Ankur Parikh; Xuezhi Wang; Sebastian Gehrmann; Manaal Faruqui; Bhuwan Dhingra; Diyi Yang; Dipanjan Das; | |
90 | ENT-DESC: Entity Description Generation By Exploring Knowledge Graph Highlight: In this paper, we introduce a large-scale and challenging dataset to facilitate the study of such a practical scenario in KG-to-text. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Liying Cheng; Dekun Wu; Lidong Bing; Yan Zhang; Zhanming Jie; Wei Lu; Luo Si; | |
91 | Small But Mighty: New Benchmarks For Split And Rephrase Highlight: We find that the widely used benchmark dataset universally contains easily exploitable syntactic cues caused by its automatic generation process. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Li Zhang; Huaiyu Zhu; Siddhartha Brahma; Yunyao Li; | |
92 | Online Back-Parsing For AMR-to-Text Generation Highlight: We propose a decoder that back predicts projected AMR graphs on the target sentence during text generation. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Xuefeng Bai; Linfeng Song; Yue Zhang; | |
93 | Reading Between The Lines: Exploring Infilling In Visual Narratives Highlight: In this paper, we tackle this problem by using infilling techniques involving prediction of missing steps in a narrative while generating textual descriptions from a sequence of images. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Khyathi Raghavi Chandu; Ruo-Ping Dong; Alan W Black; | code |
94 | Acrostic Poem Generation Highlight: We propose a new task in the area of computational creativity: acrostic poem generation in English. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Rajat Agarwal; Katharina Kann; | |
95 | Local Additivity Based Data Augmentation For Semi-supervised NER Highlight: In this work, to alleviate the dependence on labeled data, we propose a Local Additivity based Data Augmentation (LADA) method for semi-supervised NER, in which we create virtual samples by interpolating sequences close to each other. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Jiaao Chen; Zhenghui Wang; Ran Tian; Zichao Yang; Diyi Yang; | code |
96 | Grounded Compositional Outputs For Adaptive Language Modeling Highlight: In this work, we go one step beyond and propose a fully compositional output embedding layer for language models, which is further grounded in information from a structured lexicon (WordNet), namely semantically related words and free-text definitions. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Nikolaos Pappas; Phoebe Mulcaire; Noah A. Smith; | |
97 | SSMBA: Self-Supervised Manifold Based Data Augmentation For Improving Out-of-Domain Robustness Highlight: We introduce SSMBA, a data augmentation method for generating synthetic training examples by using a pair of corruption and reconstruction functions to move randomly on a data manifold. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Nathan Ng; Kyunghyun Cho; Marzyeh Ghassemi; | |
98 | SetConv: A New Approach For Learning From Imbalanced Data Highlight: To address this problem, we propose a set convolution (SetConv) operation and an episodic training strategy to extract a single representative for each class, so that classifiers can later be trained on a balanced class distribution. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Yang Gao; Yi-Fan Li; Yu Lin; Charu Aggarwal; Latifur Khan; | |
99 | Scalable Multi-Hop Relational Reasoning For Knowledge-Aware Question Answering Highlight: In this paper, we propose a novel knowledge-aware approach that equips pre-trained language models (PTLMs) has with a multi-hop relational reasoning module, named multi-hop graph relation network (MHGRN). Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Yanlin Feng; Xinyue Chen; Bill Yuchen Lin; Peifeng Wang; Jun Yan; Xiang Ren; | |
100 | Improving Bilingual Lexicon Induction For Low Frequency Words Highlight: This paper designs a Monolingual Lexicon Induction task and observes that two factors accompany the degraded accuracy of bilingual lexicon induction for rare words. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Jiaji Huang; Xingyu Cai; Kenneth Church; | |
101 | Learning VAE-LDA Models With Rounded Reparameterization Trick Highlight: In this work, we propose a new method, which we call Rounded Reparameterization Trick (RRT), to reparameterize Dirichlet distributions for the learning of VAE-LDA models. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Runzhi Tian; Yongyi Mao; Richong Zhang; | |
102 | Calibrated Language Model Fine-Tuning For In- And Out-of-Distribution Data Highlight: To mitigate this issue, we propose a regularized fine-tuning method. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Lingkai Kong; Haoming Jiang; Yuchen Zhuang; Jie Lyu; Tuo Zhao; Chao Zhang; | code |
103 | Scaling Hidden Markov Language Models Highlight: We propose methods for scaling HMMs to massive state spaces while maintaining efficient exact inference, a compact parameterization, and effective regularization. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Justin Chiu; Alexander Rush; | |
104 | Coding Textual Inputs Boosts The Accuracy Of Neural Networks Highlight: As alternatives to a text representation, we introduce Soundex, MetaPhone, NYSIIS, logogram to NLP, and develop fixed-output-length coding and its extension using Huffman coding. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Abdul Rafae Khan; Jia Xu; Weiwei Sun; | code |
105 | Learning From Task Descriptions Highlight: To take a step toward closing this gap, we introduce a framework for developing NLP systems that solve new tasks after reading their descriptions, synthesizing prior work in this area. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Orion Weller; Nicholas Lourie; Matt Gardner; Matthew Peters; | |
106 | Hashtags, Emotions, And Comments: A Large-Scale Dataset To Understand Fine-Grained Social Emotions To Online Topics Highlight: This paper studies social emotions to online discussion topics. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Keyang Ding; Jing Li; Yuji Zhang; | |
107 | Named Entity Recognition For Social Media Texts With Semantic Augmentation Highlight: In this paper, we propose a neural-based approach to NER for social media texts where both local (from running text) and augmented semantics are taken into account. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Yuyang Nie; Yuanhe Tian; Xiang Wan; Yan Song; Bo Dai; | |
108 | Coupled Hierarchical Transformer For Stance-Aware Rumor Verification In Social Media Conversations Highlight: Therefore, in this paper, to extend BERT to obtain thread representations, we first propose a Hierarchical Transformer, which divides each long thread into shorter subthreads, and employs BERT to separately represent each subthread, followed by a global Transformer layer to encode all the subthreads. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Jianfei Yu; Jing Jiang; Ling Min Serena Khoo; Hai Leong Chieu; Rui Xia; | |
109 | Social Media Attributions In The Context Of Water Crisis Highlight: In this paper, we explore the viability of using unstructured, noisy social media data to complement traditional surveys through automatically extracting attribution factors. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Rupak Sarkar; Sayantan Mahinder; Hirak Sarkar; Ashiqur KhudaBukhsh; | |
110 | On The Reliability And Validity Of Detecting Approval Of Political Actors In Tweets Highlight: In this work, we attempt to gauge the efficacy of untargeted sentiment, targeted sentiment, and stance detection methods in labeling various political actors’ approval by benchmarking them across several datasets. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Indira Sen; Fabian Flöck; Claudia Wagner; | |
111 | Towards Medical Machine Reading Comprehension With Structural Knowledge And Plain Text Highlight: As an effort, we first collect a large scale medical multi-choice question dataset (more than 21k instances) for the National Licensed Pharmacist Examination in China. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Dongfang Li; Baotian Hu; Qingcai Chen; Weihua Peng; Anqi Wang; | |
112 | Generating Radiology Reports Via Memory-driven Transformer Highlight: In this paper, we propose to generate radiology reports with memory-driven Transformer, where a relational memory is designed to record key information of the generation process and a memory-driven conditional layer normalization is applied to incorporating the memory into the decoder of Transformer. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Zhihong Chen; Yan Song; Tsung-Hui Chang; Xiang Wan; | |
113 | Planning And Generating Natural And Diverse Disfluent Texts As Augmentation For Disfluency Detection Highlight: In this work, we propose a simple Planner-Generator based disfluency generation model to generate natural and diverse disfluent texts as augmented data, where the Planner decides on where to insert disfluent segments and the Generator follows the prediction to generate corresponding disfluent segments. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Jingfeng Yang; Diyi Yang; Zhaoran Ma; | |
114 | Predicting Clinical Trial Results By Implicit Evidence Integration Highlight: To optimize the design of clinical trials, we introduce a novel Clinical Trial Result Prediction (CTRP) task. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Qiao Jin; Chuanqi Tan; Mosha Chen; Xiaozhong Liu; Songfang Huang; | |
115 | Explainable Clinical Decision Support From Text Highlight: We propose a hierarchical CNN-transformer model with explicit attention as an interpretable, multi-task clinical language model, which achieves an AUROC of 0.75 and 0.78 on sepsis and mortality prediction, respectively. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Jinyue Feng; Chantal Shaib; Frank Rudzicz; | |
116 | A Knowledge-driven Generative Model For Multi-implication Chinese Medical Procedure Entity Normalization Highlight: In this paper, we focus on Chinese medical procedure entity normalization. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Jinghui Yan; Yining Wang; Lu Xiang; Yu Zhou; Chengqing Zong; | |
117 | Combining Automatic Labelers And Expert Annotations For Accurate Radiology Report Labeling Using BERT Highlight: In this work, we introduce a BERT-based approach to medical image report labeling that exploits both the scale of available rule-based systems and the quality of expert annotations. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Akshay Smit; Saahil Jain; Pranav Rajpurkar; Anuj Pareek; Andrew Ng; Matthew Lungren; | |
118 | Benchmarking Meaning Representations In Neural Semantic Parsing Highlight: Upon identifying these gaps, we propose , a new unified benchmark on meaning representations, by integrating existing semantic parsing datasets, completing the missing logical forms, and implementing the missing execution engines. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Jiaqi Guo; Qian Liu; Jian-Guang Lou; Zhenwen Li; Xueqing Liu; Tao Xie; Ting Liu; | code |
119 | Analogous Process Structure Induction For Sub-event Sequence Prediction Highlight: In this paper, we propose an Analogous Process Structure Induction (APSI) framework, which leverages analogies among processes and conceptualization of sub-event instances to predict the whole sub-event sequence of previously unseen open-domain processes. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Hongming Zhang; Muhao Chen; Haoyu Wang; Yangqiu Song; Dan Roth; | |
120 | SLM: Learning A Discourse Language Representation With Sentence Unshuffling Highlight: We introduce Sentence-level Language Modeling, a new pre-training objective for learning a discourse language representation in a fully self-supervised manner. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Haejun Lee; Drew A. Hudson; Kangwook Lee; Christopher D. Manning; | |
121 | Detecting Fine-Grained Cross-Lingual Semantic Divergences Without Supervision By Learning To Rank Highlight: We introduce a training strategy for multilingual BERT models by learning to rank synthetic divergent examples of varying granularity. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Eleftheria Briakou; Marine Carpuat; | |
122 | A Bilingual Generative Transformer For Semantic Sentence Embedding Highlight: We propose a deep latent variable model that attempts to perform source separation on parallel sentences, isolating what they have in common in a latent semantic vector, and explaining what is left over with language-specific latent vectors. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
John Wieting; Graham Neubig; Taylor Berg-Kirkpatrick; | |
123 | Semantically Inspired AMR Alignment For The Portuguese Language Highlight: Aiming to fulfill this gap, we developed an alignment method for the Portuguese language based on a more semantically matched word-concept pair. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Rafael Anchiêta; Thiago Pardo; | |
124 | An Unsupervised Sentence Embedding Method By Mutual Information Maximization Highlight: In this paper, we propose a lightweight extension on top of BERT and a novel self-supervised learning objective based on mutual information maximization strategies to derive meaningful sentence embeddings in an unsupervised manner. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Yan Zhang; Ruidan He; Zuozhu Liu; Kwan Hui Lim; Lidong Bing; | |
125 | Compositional Phrase Alignment And Beyond Highlight: We address the phrase alignment problem by combining an unordered tree mapping algorithm and phrase representation modelling that explicitly embeds the similarity distribution in the sentences onto powerful contextualized representations. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Yuki Arase; Jun’ichi Tsujii; | |
126 | Table Fact Verification With Structure-Aware Transformer Highlight: To better utilize pre-trained transformers for table representation, we propose a Structure-Aware Transformer (SAT), which injects the table structural information into the mask of the self-attention layer. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Hongzhi Zhang; Yingyao Wang; Sirui Wang; Xuezhi Cao; Fuzheng Zhang; Zhongyuan Wang; | |
127 | Double Graph Based Reasoning For Document-level Relation Extraction Highlight: In this paper, we propose Graph Aggregation-and-Inference Network (GAIN), a method to recognize such relations for long paragraphs. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Shuang Zeng; Runxin Xu; Baobao Chang; Lei Li; | code |
128 | Event Extraction As Machine Reading Comprehension Highlight: In this paper, we propose a new learning paradigm of EE, by explicitly casting it as a machine reading comprehension problem (MRC). Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Jian Liu; Yubo Chen; Kang Liu; Wei Bi; Xiaojiang Liu; | |
129 | MAVEN: A Massive General Domain Event Detection Dataset Highlight: To alleviate these problems, we present a MAssive eVENt detection dataset (MAVEN), which contains 4,480 Wikipedia documents, 118,732 event mention instances, and 168 event types. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Xiaozhi Wang; Ziqi Wang; Xu Han; Wangyi Jiang; Rong Han; Zhiyuan Liu; Juanzi Li; Peng Li; Yankai Lin; Jie Zhou; | code |
130 | Knowledge Graph Alignment With Entity-Pair Embedding Highlight: In this work, we present a new approach that directly learns embeddings of entity-pairs for KG alignment. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Zhichun Wang; Jinjian Yang; Xiaoju Ye; | |
131 | Adaptive Attentional Network For Few-Shot Knowledge Graph Completion Highlight: This work proposes an adaptive attentional network for few-shot KG completion by learning adaptive entity and reference representations. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Jiawei Sheng; Shu Guo; Zhenyu Chen; Juwei Yue; Lihong Wang; Tingwen Liu; Hongbo Xu; | code |
132 | Pre-training Entity Relation Encoder With Intra-span And Inter-span Information Highlight: In this paper, we integrate span-related information into pre-trained encoder for entity relation extraction task. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Yijun Wang; Changzhi Sun; Yuanbin Wu; Junchi Yan; Peng Gao; Guotong Xie; | |
133 | Two Are Better Than One: Joint Entity And Relation Extraction With Table-Sequence Encoders Highlight: In this work, we propose the novel table-sequence encoders where two different encoders – a table encoder and a sequence encoder are designed to help each other in the representation learning process. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Jue Wang; Wei Lu; | |
134 | Beyond [CLS] Through Ranking By Generation Highlight: In this work, we revisit the generative framework for information retrieval and show that our generative approaches are as effective as state-of-the-art semantic similarity-based discriminative models for the answer selection task. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Cicero Nogueira dos Santos; Xiaofei Ma; Ramesh Nallapati; Zhiheng Huang; Bing Xiang; | |
135 | Tired Of Topic Models? Clusters Of Pretrained Word Embeddings Make For Fast And Good Topics Too! Highlight: The dominant approach is to use probabilistic topic models that posit a generative story, but in this paper we propose an alternative way to obtain topics: clustering pre-trained word embeddings while incorporating document information for weighted clustering and reranking top words. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Suzanna Sia; Ayush Dalmia; Sabrina J. Mielke; | |
136 | Multi-document Summarization With Maximal Marginal Relevance-guided Reinforcement Learning Highlight: To close the gap, we present RL-MMR, Maximal Margin Relevance-guided Reinforcement Learning for MDS, which unifies advanced neural SDS methods and statistical measures used in classical MDS. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Yuning Mao; Yanru Qu; Yiqing Xie; Xiang Ren; Jiawei Han; | |
137 | Improving Neural Topic Models Using Knowledge Distillation Highlight: We use knowledge distillation to combine the best attributes of probabilistic topic models and pretrained transformers. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Alexander Miserlis Hoyle; Pranav Goel; Philip Resnik; | |
138 | Short Text Topic Modeling With Topic Distribution Quantization And Negative Sampling Decoder Highlight: In this paper, to address this issue, we propose a novel neural topic model in the framework of autoencoding with a new topic distribution quantization approach generating peakier distributions that are more appropriate for modeling short texts. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Xiaobao Wu; Chunping Li; Yan Zhu; Yishu Miao; | |
139 | Querying Across Genres For Medical Claims In News Highlight: We present a query-based biomedical information retrieval task across two vastly different genres – newswire and research literature – where the goal is to find the research publication that supports the primary claim made in a health-related news article. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Chaoyuan Zuo; Narayan Acharya; Ritwik Banerjee; | |
140 | Incorporating Multimodal Information In Open-Domain Web Keyphrase Extraction Highlight: In this work, we propose a modeling approach that leverages these multi-modal signals to aid in the KPE task. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Yansen Wang; Zhen Fan; Carolyn Rose; | |
141 | CMU-MOSEAS: A Multimodal Language Dataset For Spanish, Portuguese, German And French Highlight: As a step towards building more equitable and inclusive multimodal systems, we introduce the first large-scale multimodal language dataset for Spanish, Portuguese, German and French. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
AmirAli Bagher Zadeh; Yansheng Cao; Simon Hessner; Paul Pu Liang; Soujanya Poria; Louis-Philippe Morency; | |
142 | Combining Self-Training And Self-Supervised Learning For Unsupervised Disfluency Detection Highlight: In this work, we explore the unsupervised learning paradigm which can potentially work with unlabeled text corpora that are cheaper and easier to obtain. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Shaolei Wang; Zhongyuan Wang; Wanxiang Che; Ting Liu; | |
143 | Multimodal Routing: Improving Local And Global Interpretability Of Multimodal Language Analysis Highlight: In this paper we propose, which dynamically adjusts weights between input modalities and output representations differently for each input sample. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Yao-Hung Hubert Tsai; Martin Ma; Muqiao Yang; Ruslan Salakhutdinov; Louis-Philippe Morency; | |
144 | Multistage Fusion With Forget Gate For Multimodal Summarization In Open-Domain Videos Highlight: To address these two issues, we propose a multistage fusion network with the fusion forget gate module, which builds upon this approach by modeling fine-grained interactions between the modalities through a multistep fusion schema and controlling the flow of redundant information between multimodal long sequences via a forgetting module. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Nayu Liu; Xian Sun; Hongfeng Yu; Wenkai Zhang; Guangluan Xu; | |
145 | BiST: Bi-directional Spatio-Temporal Reasoning For Video-Grounded Dialogues Highlight: To address this drawback, we proposed Bi-directional Spatio-Temporal Learning (BiST), a vision-language neural framework for high-resolution queries in videos based on textual cues. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Hung Le; Doyen Sahoo; Nancy Chen; Steven C.H. Hoi; | |
146 | UniConv: A Unified Conversational Neural Architecture For Multi-domain Task-oriented Dialogues Highlight: Unlike the existing approaches that are often designed to train each module separately, we propose UniConv – a novel unified neural architecture for end-to-end conversational systems in multi-domain task-oriented dialogues, which is designed to jointly train (i) a Bi-level State Tracker which tracks dialogue states by learning signals at both slot and domain level independently, and (ii) a Joint Dialogue Act and Response Generator which incorporates information from various input components and models dialogue acts and target responses simultaneously. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Hung Le; Doyen Sahoo; Chenghao Liu; Nancy Chen; Steven C.H. Hoi; | |
147 | GraphDialog: Integrating Graph Knowledge Into End-to-End Task-Oriented Dialogue Systems Highlight: In this paper, we address these two challenges by exploiting the graph structural information in the knowledge base and in the dependency parsing tree of the dialogue. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Shiquan Yang; Rui Zhang; Sarah Erfani; | |
148 | Structured Attention For Unsupervised Dialogue Structure Induction Highlight: In this work, we propose to incorporate structured attention layers into a Variational Recurrent Neural Network (VRNN) model with discrete latent states to learn dialogue structure in an unsupervised fashion. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Liang Qiu; Yizhou Zhao; Weiyan Shi; Yuan Liang; Feng Shi; Tao Yuan; Zhou Yu; Song-Chun Zhu; | |
149 | Cross Copy Network For Dialogue Generation Highlight: In this paper, we propose a novel network architecture – Cross Copy Networks (CCN) to explore the current dialog context and similar dialogue instances’ logical structure simultaneously. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Changzhen Ji; Xin Zhou; Yating Zhang; Xiaozhong Liu; Changlong Sun; Conghui Zhu; Tiejun Zhao; | |
150 | Multi-turn Response Selection Using Dialogue Dependency Relations Highlight: In this paper, we propose a dialogue extraction algorithm to transform a dialogue history into threads based on their dependency relations. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Qi Jia; Yizhu Liu; Siyu Ren; Kenny Zhu; Haifeng Tang; | |
151 | Parallel Interactive Networks For Multi-Domain Dialogue State Generation Highlight: In this study, we argue that the incorporation of these dependencies is crucial for the design of MDST and propose Parallel Interactive Networks (PIN) to model these dependencies. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Junfan Chen; Richong Zhang; Yongyi Mao; Jie Xu; | |
152 | SlotRefine: A Fast Non-Autoregressive Model For Joint Intent Detection And Slot Filling Highlight: In this paper, we propose a novel non-autoregressive model named SlotRefine for joint intent detection and slot filling. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Di Wu; Liang Ding; Fan Lu; Jian Xie; | |
153 | An Information Bottleneck Approach For Controlling Conciseness In Rationale Extraction Highlight: In this paper, we show that it is possible to better manage the trade-off between concise explanations and high task accuracy by optimizing a bound on the Information Bottleneck (IB) objective. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Bhargavi Paranjape; Mandar Joshi; John Thickstun; Hannaneh Hajishirzi; Luke Zettlemoyer; | |
154 | CrowS-Pairs: A Challenge Dataset For Measuring Social Biases In Masked Language Models Highlight: To measure some forms of social bias in language models against protected demographic groups in the US, we introduce the Crowdsourced Stereotype Pairs benchmark (CrowS-Pairs). Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Nikita Nangia; Clara Vania; Rasika Bhalerao; Samuel R. Bowman; | |
155 | LOGAN: Local Group Bias Detection By Clustering Highlight: To analyze and detect such local bias, we propose LOGAN, a new bias detection technique based on clustering. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Jieyu Zhao; Kai-Wei Chang; | |
156 | RNNs Can Generate Bounded Hierarchical Languages With Optimal Memory Highlight: We introduce Dyck-$(k,m)$, the language of well-nested brackets (of $k$ types) and $m$-bounded nesting depth, reflecting the bounded memory needs and long-distance dependencies of natural language syntax. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
John Hewitt; Michael Hahn; Surya Ganguli; Percy Liang; Christopher D. Manning; | |
157 | Detecting Independent Pronoun Bias With Partially-Synthetic Data Generation Highlight: We introduce a new technique for measuring bias in models, using Bayesian approximations to generate partially-synthetic data from the model itself. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Robert Munro; Alex (Carmen) Morrison; | |
158 | Visually Grounded Continual Learning Of Compositional Phrases Highlight: To study this human-like language acquisition ability, we present VisCOLL, a visually grounded language learning task, which simulates the continual acquisition of compositional phrases from streaming visual scenes. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Xisen Jin; Junyi Du; Arka Sadhu; Ram Nevatia; Xiang Ren; | |
159 | MAF: Multimodal Alignment Framework For Weakly-Supervised Phrase Grounding Highlight: Given difficulties in annotating phrase-to-object datasets at scale, we develop a Multimodal Alignment Framework (MAF) to leverage more widely-available caption-image datasets, which can then be used as a form of weak supervision. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Qinxin Wang; Hao Tan; Sheng Shen; Michael Mahoney; Zhewei Yao; | |
160 | Domain-Specific Lexical Grounding In Noisy Visual-Textual Documents Highlight: We present a simple unsupervised clustering-based method that increases precision and recall beyond object detection and image tagging baselines when evaluated on labeled subsets of the dataset. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Gregory Yauney; Jack Hessel; David Mimno; | |
161 | HERO: Hierarchical Encoder For Video+Language Omni-representation Pre-training Highlight: We present HERO, a novel framework for large-scale video+language omni-representation learning. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Linjie Li; Yen-Chun Chen; Yu Cheng; Zhe Gan; Licheng Yu; Jingjing Liu; | |
162 | Vokenization: Improving Language Understanding With Contextualized, Visual-Grounded Supervision Highlight: Therefore, we develop a technique named vokenization that extrapolates multimodal alignments to language-only data by contextually mapping language tokens to their related images (which we call vokens). Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Hao Tan; Mohit Bansal; | |
163 | Detecting Cross-Modal Inconsistency To Defend Against Neural Fake News Highlight: In this paper, we introduce the more realistic and challenging task of defending against machine-generated news that also includes images and captions. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Reuben Tan; Bryan Plummer; Kate Saenko; | |
164 | Enhancing Aspect Term Extraction With Soft Prototypes Highlight: In this paper, we propose to tackle this problem by correlating words with each other through soft prototypes. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Zhuang Chen; Tieyun Qian; | |
165 | FedED: Federated Learning Via Ensemble Distillation For Medical Relation Extraction Highlight: In this paper, we propose a privacy-preserving medical relation extraction model based on federated learning, which enables training a central model with no single piece of private local data being shared or exchanged. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Dianbo Sui; Yubo Chen; Jun Zhao; Yantao Jia; Yuantao Xie; Weijian Sun; | |
166 | Multimodal Joint Attribute Prediction And Value Extraction For E-commerce Product Highlight: In this paper, we propose a multimodal method to jointly predict product attributes and extract values from textual product descriptions with the help of the product images. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Tiangang Zhu; Yue Wang; Haoran Li; Youzheng Wu; Xiaodong He; Bowen Zhou; | code |
167 | A Predicate-Function-Argument Annotation Of Natural Language For Open-Domain Information EXpression Highlight: This paper proposes a new pipeline to build OIE systems, where an Open-domain Information eXpression (OIX) task is proposed to provide a platform for all OIE strategies. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Mingming Sun; Wenyue Hua; Zoey Liu; Xin Wang; Kangjie Zheng; Ping Li; | |
168 | Retrofitting Structure-aware Transformer Language Model For End Tasks Highlight: We consider retrofitting structure-aware Transformer language model for facilitating end tasks by proposing to exploit syntactic distance to encode both the phrasal constituency and dependency connection into the language model. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Hao Fei; Yafeng Ren; Donghong Ji; | |
169 | Lightweight, Dynamic Graph Convolutional Networks For AMR-to-Text Generation Highlight: In this paper, we introduce a dynamic fusion mechanism, proposing Lightweight Dynamic Graph Convolutional Networks (LDGCNs) that capture richer non-local interactions by synthesizing higher order information from the input graphs. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Yan Zhang; Zhijiang Guo; Zhiyang Teng; Wei Lu; Shay B. Cohen; Zuozhu Liu; Lidong Bing; | |
170 | If Beam Search Is The Answer, What Was The Question? Highlight: We frame beam search as the exact solution to a different decoding objective in order to gain insights into why high probability under a model alone may not indicate adequacy. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Clara Meister; Ryan Cotterell; Tim Vieira; | |
171 | Understanding The Mechanics Of SPIGOT: Surrogate Gradients For Latent Structure Learning Highlight: In this paper, we focus on surrogate gradients, a popular strategy to deal with this problem. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Tsvetomila Mihaylova; Vlad Niculae; André F. T. Martins; | |
172 | Is The Best Better? Bayesian Statistical Model Comparison For Natural Language Processing Highlight: We propose a Bayesian statistical model comparison technique which uses k-fold cross-validation across multiple data sets to estimate the likelihood that one model will outperform the other, or that the two will produce practically equivalent results. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Piotr Szymański; Kyle Gorman; | |
173 | Exploring Logically Dependent Multi-task Learning With Causal Inference Highlight: In this paper, we view logically dependent MTL from the perspective of causal inference and suggest a mediation assumption instead of the confounding assumption in conventional MTL models. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Wenqing Chen; Jidong Tian; Liqiang Xiao; Hao He; Yaohui Jin; | |
174 | Masking As An Efficient Alternative To Finetuning For Pretrained Language Models Highlight: We present an efficient method of utilizing pretrained language models, where we learn selective binary masks for pretrained weights in lieu of modifying them through finetuning. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Mengjie Zhao; Tao Lin; Fei Mi; Martin Jaggi; Hinrich Schütze; | |
175 | Dynamic Context Selection For Document-level Neural Machine Translation Via Reinforcement Learning Highlight: To address this problem, we propose an effective approach to select dynamic context so that the document-level translation model can utilize the more useful selected context sentences to produce better translations. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Xiaomian Kang; Yang Zhao; Jiajun Zhang; Chengqing Zong; | |
176 | Data Rejuvenation: Exploiting Inactive Training Examples For Neural Machine Translation Highlight: In this work, we explore to identify the inactive training examples which contribute less to the model performance, and show that the existence of inactive examples depends on the data distribution. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Wenxiang Jiao; Xing Wang; Shilin He; Irwin King; Michael Lyu; Zhaopeng Tu; | |
177 | Pronoun-Targeted Fine-tuning For NMT With Hybrid Losses Highlight: We introduce a class of conditional generative-discriminative hybrid losses that we use to fine-tune a trained machine translation model. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Prathyusha Jwalapuram; Shafiq Joty; Youlin Shen; | |
178 | Learning Adaptive Segmentation Policy For Simultaneous Translation Highlight: Inspired by human interpreters, we propose a novel adaptive segmentation policy for simultaneous translation. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Ruiqing Zhang; Chuanqiang Zhang; Zhongjun He; Hua Wu; Haifeng Wang; | |
179 | Learn To Cross-lingual Transfer With Meta Graph Learning Across Heterogeneous Languages Highlight: To address the issues, we propose a meta graph learning (MGL) method. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Zheng Li; Mukul Kumar; William Headden; Bing Yin; Ying Wei; Yu Zhang; Qiang Yang; | |
180 | UDapter: Language Adaptation For Truly Universal Dependency Parsing Highlight: To address this, we propose a novel multilingual task adaptation approach based on contextual parameter generation and adapter modules. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Ahmet Üstün; Arianna Bisazza; Gosse Bouma; Gertjan van Noord; | |
181 | Uncertainty-Aware Label Refinement For Sequence Labeling Highlight: In this work, we introduce a novel two-stage label decoding framework to model long-term label dependencies, while being much more computationally efficient. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Tao Gui; Jiacheng Ye; Qi Zhang; Zhengyan Li; Zichu Fei; Yeyun Gong; Xuanjing Huang; | |
182 | Adversarial Attack And Defense Of Structured Prediction Models Highlight: In this paper, we investigate attacks and defenses for structured prediction tasks in NLP. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Wenjuan Han; Liwen Zhang; Yong Jiang; Kewei Tu; | |
183 | Position-Aware Tagging For Aspect Sentiment Triplet Extraction Highlight: In this work, we propose the first end-to-end model with a novel position-aware tagging scheme that is capable of jointly extracting the triplets. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Lu Xu; Hao Li; Wei Lu; Lidong Bing; | |
184 | Simultaneous Machine Translation With Visual Context Highlight: In this paper, we seek to understand whether the addition of visual information can compensate for the missing source context. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Ozan Caglayan; Julia Ive; Veneta Haralampieva; Pranava Madhyastha; Loïc Barrault; Lucia Specia; | |
185 | XCOPA: A Multilingual Dataset For Causal Commonsense Reasoning Highlight: Motivated by both demands, we introduce Cross-lingual Choice of Plausible Alternatives (XCOPA), a typologically diverse multilingual dataset for causal commonsense reasoning in 11 languages, which includes resource-poor languages like Eastern Apur{\’\i}mac Quechua and Haitian Creole. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Edoardo Maria Ponti; Goran Glavaš; Olga Majewska; Qianchu Liu; Ivan Vulić; Anna Korhonen; | |
186 | The Secret Is In The Spectra: Predicting Cross-lingual Task Performance With Spectral Similarity Measures Highlight: In this work we present a large-scale study focused on the correlations between monolingual embedding space similarity and task performance, covering thousands of language pairs and four different tasks: BLI, parsing, POS tagging and MT. We hypothesize that statistics of the spectrum of each monolingual embedding space indicate how well they can be aligned. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Haim Dubossarsky; Ivan Vulić; Roi Reichart; Anna Korhonen; | |
187 | Bridging Linguistic Typology And Multilingual Machine Translation With Multi-View Language Representations Highlight: We propose to fuse both views using singular vector canonical correlation analysis and study what kind of information is induced from each source. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Arturo Oncevay; Barry Haddow; Alexandra Birch; | |
188 | AnswerFact: Fact Checking In Product Question Answering Highlight: To tackle this issue, we investigate to predict the veracity of answers in this paper and introduce AnswerFact, a large scale fact checking dataset from product question answering forums. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Wenxuan Zhang; Yang Deng; Jing Ma; Wai Lam; | |
189 | Context-Aware Answer Extraction In Question Answering Highlight: To resolve this issue, we propose BLANC (BLock AttentioN for Context prediction) based on two main ideas: context prediction as an auxiliary task in multi-task learning manner, and a block attention method that learns the context prediction task. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Yeon Seonwoo; Ji-Hoon Kim; Jung-Woo Ha; Alice Oh; | |
190 | What Do Models Learn From Question Answering Datasets? Highlight: In this paper, we investigate if models are learning reading comprehension from QA datasets by evaluating BERT-based models across five datasets. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Priyanka Sen; Amir Saffari; | code |
191 | Discern: Discourse-Aware Entailment Reasoning Network For Conversational Machine Reading Highlight: In this work, we propose Discern, a discourse-aware entailment reasoning network to strengthen the connection and enhance the understanding of both document and dialog. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Yifan Gao; Chien-Sheng Wu; Jingjing Li; Shafiq Joty; Steven C.H. Hoi; Caiming Xiong; Irwin King; Michael Lyu; | code |
192 | A Method For Building A Commonsense Inference Dataset Based On Basic Events Highlight: We present a scalable, low-bias, and low-cost method for building a commonsense inference dataset that combines automatic extraction from a corpus and crowdsourcing. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Kazumasa Omura; Daisuke Kawahara; Sadao Kurohashi; | |
193 | Neural Deepfake Detection With Factual Structure Of Text Highlight: To address this, we propose a graph-based model that utilizes the factual structure of a document for deepfake detection of text. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Wanjun Zhong; Duyu Tang; Zenan Xu; Ruize Wang; Nan Duan; Ming Zhou; Jiahai Wang; Jian Yin; | |
194 | MultiCQA: Zero-Shot Transfer Of Self-Supervised Text Matching Models On A Massive Scale Highlight: We propose to incorporate self-supervised with supervised multi-task learning on all available source domains. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Andreas Rücklé; Jonas Pfeiffer; Iryna Gurevych; | |
195 | XL-AMR: Enabling Cross-Lingual AMR Parsing With Transfer Learning Techniques Highlight: In this work we tackle these two problems so as to enable cross-lingual AMR parsing: we explore different transfer learning techniques for producing automatic AMR annotations across languages and develop a cross-lingual AMR parser, XL-AMR. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Rexhina Blloshmi; Rocco Tripodi; Roberto Navigli; | |
196 | Improving AMR Parsing With Sequence-to-Sequence Pre-training Highlight: In this paper, we focus on sequence-to-sequence (seq2seq) AMR parsing and propose a seq2seq pre-training approach to build pre-trained models in both single and joint way on three relevant tasks, i.e., machine translation, syntactic parsing, and AMR parsing itself. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Dongqin Xu; Junhui Li; Muhua Zhu; Min Zhang; Guodong Zhou; | code |
197 | Hate-Speech And Offensive Language Detection In Roman Urdu Highlight: In this study, we: (1) Present a lexicon of hateful words in RU, (2) Develop an annotated dataset called RUHSOLD consisting of 10,012 tweets in RU with both coarse-grained and fine-grained labels of hate-speech and offensive language, (3) Explore the feasibility of transfer learning of five existing embedding models to RU, (4) Propose a novel deep learning architecture called CNN-gram for hate-speech and offensive language detection and compare its performance with seven current baseline approaches on RUHSOLD dataset, and (5) Train domain-specific embeddings on more than 4.7 million tweets and make them publicly available. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Hammad Rizwan; Muhammad Haroon Shakeel; Asim Karim; | |
198 | Suicidal Risk Detection For Military Personnel Highlight: We analyze social media for detecting the suicidal risk of military personnel, which is especially crucial for countries with compulsory military service such as the Republic of Korea. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Sungjoon Park; Kiwoong Park; Jaimeen Ahn; Alice Oh; | |
199 | Comparative Evaluation Of Label-Agnostic Selection Bias In Multilingual Hate Speech Datasets Highlight: We examine selection bias in hate speech in a language and label independent fashion. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Nedjma Ousidhoum; Yangqiu Song; Dit-Yan Yeung; | |
200 | HENIN: Learning Heterogeneous Neural Interaction Networks For Explainable Cyberbullying Detection On Social Media Highlight: In this paper, therefore, we propose a novel deep model, HEterogeneous Neural Interaction Networks (HENIN), for explainable cyberbullying detection. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Hsin-Yu Chen; Cheng-Te Li; | |
201 | Reactive Supervision: A New Method For Collecting Sarcasm Data Highlight: We introduce reactive supervision, a novel data collection method that utilizes the dynamics of online conversations to overcome the limitations of existing data collection techniques. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Boaz Shmueli; Lun-Wei Ku; Soumya Ray; | |
202 | Self-Induced Curriculum Learning In Self-Supervised Neural Machine Translation Highlight: In this study, we provide an in-depth analysis of the sampling choices the SSNMT model makes during training. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Dana Ruiter; Josef van Genabith; Cristina España-Bonet; | |
203 | Towards Reasonably-Sized Character-Level Transformer NMT By Finetuning Subword Systems Highlight: We show that by initially training a subword model and then finetuning it on characters, we can obtain a neural machine translation model that works at the character level without requiring token segmentation. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Jindřich Libovický; Alexander Fraser; | |
204 | Transfer Learning And Distant Supervision For Multilingual Transformer Models: A Study On African Languages Highlight: In this work, we study trends in performance for different amounts of available resources for the three African languages Hausa, isiXhosa and on both NER and topic classification. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Michael A. Hedderich; David Adelani; Dawei Zhu; Jesujoba Alabi; Udia Markus; Dietrich Klakow; | |
205 | Translation Quality Estimation By Jointly Learning To Score And Rank Highlight: In order to make use of different types of human evaluation data for supervised learning, we present a multi-task learning QE model that jointly learns two tasks: score a translation and rank two translations. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Jingyi Zhang; Josef van Genabith; | |
206 | Direct Segmentation Models For Streaming Speech Translation Highlight: This work proposes novel segmentation models for streaming ST that incorporate not only textual, but also acoustic information to decide when the ASR output is split into a chunk. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Javier Iranzo-Sánchez; Adrià Giménez Pastor; Joan Albert Silvestre-Cerdà; Pau Baquero-Arnal; Jorge Civera Saiz; Alfons Juan; | |
207 | Not Low-Resource Anymore: Aligner Ensembling, Batch Filtering, And New Datasets For Bengali-English Machine Translation Highlight: In this work, we build a customized sentence segmenter for Bengali and propose two novel methods for parallel corpus creation on low-resource setups: aligner ensembling and batch filtering. We release the segmenter, parallel corpus, and the evaluation set, thus elevating Bengali from its low-resource status. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Tahmid Hasan; Abhik Bhattacharjee; Kazi Samin; Masum Hasan; Madhusudan Basak; M. Sohel Rahman; Rifat Shahriyar; | code |
208 | CSP:Code-Switching Pre-training For Neural Machine Translation Highlight: This paper proposes a new pre-training method, called Code-Switching Pre-training (CSP for short) for Neural Machine Translation (NMT). Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Zhen Yang; Bojie Hu; Ambyera Han; Shen Huang; Qi Ju; | |
209 | Type B Reflexivization As An Unambiguous Testbed For Multilingual Multi-Task Gender Bias Highlight: We present a multilingual, multi-task challenge dataset, which spans four languages and four NLP tasks and focuses only on this phenomenon. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Ana Valeria González; Maria Barrett; Rasmus Hvingelby; Kellie Webster; Anders Søgaard; | |
210 | Pre-training Multilingual Neural Machine Translation By Leveraging Alignment Information Highlight: We propose mRASP, an approach to pre-train a universal multilingual neural machine translation model. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Zehui Lin; Xiao Pan; Mingxuan Wang; Xipeng Qiu; Jiangtao Feng; Hao Zhou; Lei Li; | code |
211 | Losing Heads In The Lottery: Pruning Transformer Attention In Neural Machine Translation Highlight: In this paper, we apply the lottery ticket hypothesis to prune heads in the early stages of training. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Maximiliana Behnke; Kenneth Heafield; | |
212 | Towards Enhancing Faithfulness For Neural Machine Translation Highlight: In this paper, we propose a novel training strategy with a multi-task learning paradigm to build a faithfulness enhanced NMT model (named FEnmt). Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Rongxiang Weng; Heng Yu; Xiangpeng Wei; Weihua Luo; | |
213 | COMET: A Neural Framework For MT Evaluation Highlight: We present COMET, a neural framework for training multilingual machine translation evaluation models which obtains new state-of-the-art levels of correlation with human judgements. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Ricardo Rei; Craig Stewart; Ana C Farinha; Alon Lavie; | |
214 | Reusing A Pretrained Language Model On Languages With Limited Corpora For Unsupervised NMT Highlight: We present an effective approach that reuses an LM that is pretrained only on the high-resource language. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Alexandra Chronopoulou; Dario Stojanovski; Alexander Fraser; | |
215 | LNMap: Departures From Isomorphic Assumption In Bilingual Lexicon Induction Through Non-Linear Mapping In Latent Space Highlight: In this work, we propose a novel semi-supervised method to learn cross-lingual word embeddings for BLI. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Tasnim Mohiuddin; M Saiful Bari; Shafiq Joty; | |
216 | Uncertainty-Aware Semantic Augmentation For Neural Machine Translation Highlight: To address this problem, we propose uncertainty-aware semantic augmentation, which explicitly captures the universal semantic information among multiple semantically-equivalent source sentences and enhances the hidden representations with this information for better translations. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Xiangpeng Wei; Heng Yu; Yue Hu; Rongxiang Weng; Luxi Xing; Weihua Luo; | |
217 | Can Automatic Post-Editing Improve NMT? Highlight: We hypothesize that APE models have been underperforming in improving NMT translations due to the lack of adequate supervision. To ascertain our hypothesis, we compile a larger corpus of human post-edits of English to German NMT. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Shamil Chollampatt; Raymond Hendy Susanto; Liling Tan; Ewa Szymanska; | code |
218 | Parsing Gapping Constructions Based On Grammatical And Semantic Roles Highlight: This paper proposes a method of parsing sentences with gapping to recover elided elements. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Yoshihide Kato; Shigeki Matsubara; | |
219 | Span-based Discontinuous Constituency Parsing: A Family Of Exact Chart-based Algorithms With Time Complexities From O(n^6) Down To O(n^3) Highlight: We introduce a novel chart-based algorithm for span-based parsing of discontinuous constituency trees of block degree two, including ill-nested structures. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Caio Corro; | |
220 | Some Languages Seem Easier To Parse Because Their Treebanks Leak Highlight: We compute graph isomorphisms, and show that, treebank size aside, overlap between training and test graphs explain more of the observed variation than standard explanations such as the above. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Anders Søgaard; | |
221 | Discontinuous Constituent Parsing As Sequence Labeling Highlight: This paper reduces discontinuous parsing to sequence labeling. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
David Vilares; Carlos Gómez-Rodríguez; | |
222 | Modularized Syntactic Neural Networks For Sentence Classification Highlight: This paper focuses on tree-based modeling for the sentence classification task. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Haiyan Wu; Ying Liu; Shaoyun Shi; | |
223 | TED-CDB: A Large-Scale Chinese Discourse Relation Dataset On TED Talks Highlight: As different genres are known to differ in their communicative properties and as previously, for Chinese, discourse relations have only been annotated over news text, we have created the TED-CDB dataset. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Wanqiu Long; Bonnie Webber; Deyi Xiong; | |
224 | QADiscourse – Discourse Relations As QA Pairs: Representation, Crowdsourcing And Baselines Highlight: This paper proposes a novel representation of discourse relations as QA pairs, which in turn allows us to crowd-source wide-coverage data annotated with discourse relations, via an intuitively appealing interface for composing such questions and answers. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Valentina Pyatkin; Ayal Klein; Reut Tsarfaty; Ido Dagan; | |
225 | Discourse Self-Attention For Discourse Element Identification In Argumentative Student Essays Highlight: This paper proposes to adapt self-attention to discourse level for modeling discourse elements in argumentative student essays. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Wei Song; Ziyao Song; Ruiji Fu; Lizhen Liu; Miaomiao Cheng; Ting Liu; | |
226 | MEGATRON-CNTRL: Controllable Story Generation With External Knowledge Using Large-Scale Language Models Highlight: In this paper, we propose MEGATRON-CNTRL, a novel framework that uses large-scale language models and adds control to text generation by incorporating an external knowledge base. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Peng Xu; Mostofa Patwary; Mohammad Shoeybi; Raul Puri; Pascale Fung; Anima Anandkumar; Bryan Catanzaro; | |
227 | Incomplete Utterance Rewriting As Semantic Segmentation Highlight: In this paper, we present a novel and extensive approach, which formulates it as a semantic segmentation task. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Qian Liu; Bei Chen; Jian-Guang Lou; Bin Zhou; Dongmei Zhang; | |
228 | Improving Grammatical Error Correction Models With Purpose-Built Adversarial Examples Highlight: We propose a method inspired by adversarial training to generate more meaningful and valuable training examples by continually identifying the weak spots of a model, and to enhance the model by gradually adding the generated adversarial examples to the training set. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Lihao Wang; Xiaoqing Zheng; | |
229 | Homophonic Pun Generation With Lexically Constrained Rewriting Highlight: In this paper, we focus on the task of generating a pun sentence given a pair of homophones. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Zhiwei Yu; Hongyu Zang; Xiaojun Wan; | |
230 | How To Make Neural Natural Language Generation As Reliable As Templates In Task-Oriented Dialogue Highlight: To overcome this issue, we propose a data augmentation approach which allows us to restrict the output of a network and guarantee reliability. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Henry Elder; Alexander O’Connor; Jennifer Foster; | |
231 | Multilingual AMR-to-Text Generation Highlight: In this work, we focus on Abstract Meaning Representations (AMRs) as structured input, where previous research has overwhelmingly focused on generating only into English. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Angela Fan; Claire Gardent; | |
232 | Exploring The Linear Subspace Hypothesis In Gender Bias Mitigation Highlight: In this work, we generalize their method to a kernelized, non-linear version. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Francisco Vargas; Ryan Cotterell; | |
233 | Lifelong Language Knowledge Distillation Highlight: To address this issue, we present Lifelong Language Knowledge Distillation (L2KD), a simple but efficient method that can be easily applied to existing LLL architectures in order to mitigate the degradation. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Yung-Sung Chuang; Shang-Yu Su; Yun-Nung Chen; | |
234 | Sparse Parallel Training Of Hierarchical Dirichlet Process Topic Models Highlight: In this work, we study data-parallel training for the hierarchical Dirichlet process (HDP) topic model. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Alexander Terenin; Måns Magnusson; Leif Jonsson; | |
235 | Multi-label Few/Zero-shot Learning With Knowledge Aggregated From Multiple Label Graphs Highlight: In this paper, we present a simple multi-graph aggregation model that fuses knowledge from multiple label graphs encoding different semantic label relationships in order to study how the aggregated knowledge can benefit multi-label zero/few-shot document classification. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Jueqing Lu; Lan Du; Ming Liu; Joanna Dipnall; | |
236 | Word Rotator’s Distance Highlight: Accordingly, we propose decoupling word vectors into their norm and direction then computing the alignment-based similarity with the help of earth mover’s distance (optimal transport), which we refer to as word rotator’s distance. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Sho Yokoi; Ryo Takahashi; Reina Akama; Jun Suzuki; Kentaro Inui; | code |
237 | Disentangle-based Continual Graph Representation Learning Highlight: To address this issue, we study the problem of continual graph representation learning which aims to continually train a GE model on new data to learn incessantly emerging multi-relational data while avoiding catastrophically forgetting old learned knowledge. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Xiaoyu Kou; Yankai Lin; Shaobo Liu; Peng Li; Jie Zhou; Yan Zhang; | code |
238 | Semi-Supervised Bilingual Lexicon Induction With Two-way Interaction Highlight: In this paper, we propose a new semi-supervised BLI framework to encourage the interaction between the supervised signal and unsupervised alignment. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Xu Zhao; Zihao Wang; Hao Wu; Yong Zhang; | |
239 | Wasserstein Distance Regularized Sequence Representation For Text Matching In Asymmetrical Domains Highlight: In this paper, we propose a novel match method tailored for text matching in asymmetrical domains, called WD-Match. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Weijie Yu; Chen Xu; Jun Xu; Liang Pang; Xiaopeng Gao; Xiaozhao Wang; Ji-Rong Wen; | |
240 | A Simple Approach To Learning Unsupervised Multilingual Embeddings Highlight: In contrast, we propose a simple approach by decoupling the above two sub-problems and solving them separately, one after another, using existing techniques. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Pratik Jawanpuria; Mayank Meghwanshi; Bamdev Mishra; | |
241 | Bootstrapped Q-learning With Context Relevant Observation Pruning To Generalize In Text-based Games Highlight: To address this issue, we propose Context Relevant Episodic State Truncation (CREST) for irrelevant token removal in observation text for improved generalization. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Subhajit Chaudhury; Daiki Kimura; Kartik Talamadupula; Michiaki Tatsubori; Asim Munawar; Ryuki Tachibana; | |
242 | BERT-EMD: Many-to-Many Layer Mapping For BERT Compression With Earth Mover’s Distance Highlight: In this paper, we propose a novel BERT distillation method based on many-to-many layer mapping, which allows each intermediate student layer to learn from any intermediate teacher layers. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Jianquan Li; Xiaokang Liu; Honghong Zhao; Ruifeng Xu; Min Yang; Yaohong Jin; | |
243 | Slot Attention With Value Normalization For Multi-Domain Dialogue State Tracking Highlight: In this paper, we propose a new architecture to cleverly exploit ontology, which consists of Slot Attention (SA) and Value Normalization (VN), referred to as SAVN. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Yexiang Wang; Yi Guo; Siqi Zhu; | |
244 | Don’t Read Too Much Into It: Adaptive Computation For Open-Domain Question Answering Highlight: To reduce this cost, we propose the use of adaptive computation to control the computational budget allocated for the passages to be read. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Yuxiang Wu; Sebastian Riedel; Pasquale Minervini; Pontus Stenetorp; | |
245 | Multi-Step Inference For Reasoning Over Paragraphs Highlight: We present a middle ground between these two extremes: a compositional model reminiscent of neural module networks that can perform chained logical reasoning. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Jiangming Liu; Matt Gardner; Shay B. Cohen; Mirella Lapata; | |
246 | Learning A Cost-Effective Annotation Policy For Question Answering Highlight: As a remedy, we propose a novel framework for annotating QA datasets that entails learning a cost-effective annotation policy and a semi-supervised annotation scheme. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Bernhard Kratzwald; Stefan Feuerriegel; Huan Sun; | |
247 | Scene Restoring For Narrative Machine Reading Comprehension Highlight: Inspired by this behavior of humans, we propose a method to let the machine imagine a scene during reading narrative for better comprehension. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Zhixing Tian; Yuanzhe Zhang; Kang Liu; Jun Zhao; Yantao Jia; Zhicheng Sheng; | |
248 | A Simple And Effective Model For Answering Multi-span Questions Highlight: In this work, we propose a simple architecture for answering multi-span questions by casting the task as a sequence tagging problem, namely, predicting for each input token whether it should be part of the output or not. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Elad Segal; Avia Efrat; Mor Shoham; Amir Globerson; Jonathan Berant; | |
249 | Top-Rank-Focused Adaptive Vote Collection For The Evaluation Of Domain-Specific Semantic Models Highlight: In this work, we give a threefold contribution to address these requirements: (i) we define a protocol for the construction, based on adaptive pairwise comparisons, of a relatedness-based evaluation dataset tailored on the available resources and optimized to be particularly accurate in top-rank evaluation; (ii) we define appropriate metrics, extensions of well-known ranking correlation coefficients, to evaluate a semantic model via the aforementioned dataset by taking into account the greater significance of top ranks. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Pierangelo Lombardo; Alessio Boiardi; Luca Colombo; Angelo Schiavone; Nicolò Tamagnone; | |
250 | Meta Fine-Tuning Neural Language Models For Multi-Domain Text Mining Highlight: In this paper, we propose an effective learning procedure named Meta Fine-Tuning (MFT), serving as a meta-learner to solve a group of similar NLP tasks for neural language models. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Chengyu Wang; Minghui Qiu; Jun Huang; Xiaofeng He; | |
251 | Incorporating Behavioral Hypotheses For Query Generation Highlight: This paper induces these behavioral biases as hypotheses for query generation, where a generic encoder-decoder Transformer framework is presented to aggregate arbitrary hypotheses of choice. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Ruey-Cheng Chen; Chia-Jung Lee; | |
252 | Conditional Causal Relationships Between Emotions And Causes In Texts Highlight: To address such an issue, we propose a new task of determining whether or not an input pair of emotion and cause has a valid causal relationship under different contexts, and construct a corresponding dataset via manual annotation and negative sampling based on an existing benchmark dataset. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Xinhong Chen; Qing Li; Jianping Wang; | |
253 | COMETA: A Corpus For Medical Entity Linking In The Social Media Highlight: To address this we introduce a new corpus called COMETA, consisting of 20k English biomedical entity mentions from Reddit expert-annotated with links to SNOMED CT, a widely-used medical knowledge graph. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Marco Basaldella; Fangyu Liu; Ehsan Shareghi; Nigel Collier; | |
254 | Pareto Probing: Trading Off Accuracy For Complexity Highlight: In our contribution to this discussion, we argue, first, for a probe metric that reflects the trade-off between probe complexity and performance: the Pareto hypervolume. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Tiago Pimentel; Naomi Saphra; Adina Williams; Ryan Cotterell; | |
255 | Interpretation Of NLP Models Through Input Marginalization Highlight: In this study, we raise the out-of-distribution problem induced by the existing interpretation methods and present a remedy; we propose to marginalize each token out. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Siwon Kim; Jihun Yi; Eunji Kim; Sungroh Yoon; | |
256 | Generating Label Cohesive And Well-Formed Adversarial Claims Highlight: We extend the HotFlip attack algorithm used for universal trigger generation by jointly minimizing the target class loss of a fact checking model and the entailment class loss of an auxiliary natural language inference model. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Pepa Atanasova; Dustin Wright; Isabelle Augenstein; | code |
257 | Are All Good Word Vector Spaces Isomorphic? Highlight: In this work, we ask whether non-isomorphism is also crucially a sign of degenerate word vector spaces. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Ivan Vulić; Sebastian Ruder; Anders Søgaard; | |
258 | Cold-Start And Interpretability: Turning Regular Expressions Into Trainable Recurrent Neural Networks Highlight: In this paper, we propose a type of recurrent neural networks called FA-RNNs that combine the advantages of neural networks and regular expression rules. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Chengyue Jiang; Yinggong Zhao; Shanbo Chu; Libin Shen; Kewei Tu; | |
259 | When BERT Plays The Lottery, All Tickets Are Winning Highlight: For fine-tuned BERT, we show that (a) it is possible to find subnetworks achieving performance that is comparable with that of the full model, and (b) similarly-sized subnetworks sampled from the rest of the model perform worse. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Sai Prasanna; Anna Rogers; Anna Rumshisky; | |
260 | On The Weak Link Between Importance And Prunability Of Attention Heads Highlight: Given the success of Transformer-based models, two directions of study have emerged: interpreting role of individual attention heads and down-sizing the models for efficiency. Our work straddles these two streams: We analyse the importance of basing pruning strategies on the interpreted role of the attention heads. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Aakriti Budhraja; Madhura Pande; Preksha Nema; Pratyush Kumar; Mitesh M. Khapra; | |
261 | Towards Interpreting BERT For Reading Comprehension Based QA Highlight: In this work, we attempt to interpret BERT for RCQA. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Sahana Ramnath; Preksha Nema; Deep Sahni; Mitesh M. Khapra; | code |
262 | How Do Decisions Emerge Across Layers In Neural Models? Interpretation With Differentiable Masking Highlight: To deal with these challenges, we introduce Differentiable Masking. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Nicola De Cao; Michael Sejr Schlichtkrull; Wilker Aziz; Ivan Titov; | |
263 | A Diagnostic Study Of Explainability Techniques For Text Classification Highlight: In this paper, we develop a comprehensive list of diagnostic properties for evaluating existing explainability techniques. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Pepa Atanasova; Jakob Grue Simonsen; Christina Lioma; Isabelle Augenstein; | code |
264 | STL-CQA: Structure-based Transformers With Localization And Encoding For Chart Question Answering Highlight: We propose STL-CQA which improves the question/answering through sequential elements localization, question encoding and then, a structural transformer-based learning approach. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Hrituraj Singh; Sumit Shekhar; | |
265 | Learning To Contrast The Counterfactual Samples For Robust Visual Question Answering Highlight: Therefore, we introduce a novel self-supervised contrastive learning mechanism to learn the relationship between original samples, factual samples and counterfactual samples. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Zujie Liang; Weitao Jiang; Haifeng Hu; Jiaying Zhu; | |
266 | Learning Physical Common Sense As Knowledge Graph Completion Via BERT Data Augmentation And Constrained Tucker Factorization Highlight: In this paper, we formulate physical commonsense learning as a knowledge graph completion problem to better use the latent relationships among training samples. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Zhenjie Zhao; Evangelos Papalexakis; Xiaojuan Ma; | |
267 | A Visually-grounded First-person Dialogue Dataset With Verbal And Non-verbal Responses Highlight: In this paper, we propose a visually-grounded first-person dialogue (VFD) dataset with verbal and non-verbal responses. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Hisashi Kamezawa; Noriki Nishida; Nobuyuki Shimizu; Takashi Miyazaki; Hideki Nakayama; | |
268 | Cross-Media Keyphrase Prediction: A Unified Framework With Multi-Modality Multi-Head Attention And Image Wordings Highlight: In this work, we explore the joint effects of texts and images in predicting the keyphrases for a multimedia post. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Yue Wang; Jing Li; Michael Lyu; Irwin King; | |
269 | VD-BERT: A Unified Vision And Dialog Transformer With BERT Highlight: By contrast, in this work, we propose VD-BERT, a simple yet effective framework of unified vision-dialog Transformer that leverages the pretrained BERT language models for Visual Dialog tasks. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Yue Wang; Shafiq Joty; Michael Lyu; Irwin King; Caiming Xiong; Steven C.H. Hoi; | code |
270 | The Grammar Of Emergent Languages Highlight: In this paper, we consider the syntactic properties of languages emerged in referential games, using unsupervised grammar induction (UGI) techniques originally designed to analyse natural language. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Oskar van der Wal; Silvan de Boer; Elia Bruni; Dieuwke Hupkes; | |
271 | Sub-Instruction Aware Vision-and-Language Navigation Highlight: In this work, we focus on the granularity of the visual and language sequences as well as the traceability of agents through the completion of an instruction. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Yicong Hong; Cristian Rodriguez; Qi Wu; Stephen Gould; | code |
272 | Knowledge-Grounded Dialogue Generation With Pre-trained Language Models Highlight: To leverage the redundant external knowledge under capacity constraint, we propose equipping response generation defined by a pre-trained language model with a knowledge selection module, and an unsupervised approach to jointly optimizing knowledge selection and response generation with unlabeled dialogues. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Xueliang Zhao; Wei Wu; Can Xu; Chongyang Tao; Dongyan Zhao; Rui Yan; | |
273 | MinTL: Minimalist Transfer Learning For Task-Oriented Dialogue Systems Highlight: In this paper, we propose Minimalist Transfer Learning (MinTL) to simplify the system design process of task-oriented dialogue systems and alleviate the over-dependency on annotated data. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Zhaojiang Lin; Andrea Madotto; Genta Indra Winata; Pascale Fung; | |
274 | Variational Hierarchical Dialog Autoencoder For Dialog State Tracking Data Augmentation Highlight: In this work, we extend this approach to the task of dialog state tracking for goaloriented dialogs. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Kang Min Yoo; Hanbit Lee; Franck Dernoncourt; Trung Bui; Walter Chang; Sang-goo Lee; | |
275 | Bridging The Gap Between Prior And Posterior Knowledge Selection For Knowledge-Grounded Dialogue Generation Highlight: Here, we deal with these issues on two aspects: (1) We enhance the prior selection module with the necessary posterior information obtained from the specially designed Posterior Information Prediction Module (PIPM); (2) We propose a Knowledge Distillation Based Training Strategy (KDBTS) to train the decoder with the knowledge selected from the prior distribution, removing the exposure bias of knowledge selection. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Xiuyi Chen; Fandong Meng; Peng Li; Feilong Chen; Shuang Xu; Bo Xu; Jie Zhou; | |
276 | Counterfactual Off-Policy Training For Neural Dialogue Generation Highlight: In this paper, we propose to explore potential responses by counterfactual reasoning. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Qingfu Zhu; Wei-Nan Zhang; Ting Liu; William Yang Wang; | |
277 | Dialogue Distillation: Open-Domain Dialogue Augmentation Using Unpaired Data Highlight: To address this data dilemma, we propose a novel data augmentation method for training open-domain dialogue models by utilizing unpaired data. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Rongsheng Zhang; Yinhe Zheng; Jianzhi Shao; Xiaoxi Mao; Yadong Xi; Minlie Huang; | |
278 | Task-Completion Dialogue Policy Learning Via Monte Carlo Tree Search With Dueling Network Highlight: We introduce a framework of Monte Carlo Tree Search with Double-q Dueling network (MCTS-DDU) for task-completion dialogue policy learning. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Sihan Wang; Kaijie Zhou; Kunfeng Lai; Jianping Shen; | |
279 | Learning A Simple And Effective Model For Multi-turn Response Generation With Auxiliary Tasks Highlight: In this work, we pursue a model that has a simple structure yet can effectively leverage conversation contexts for response generation. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Yufan Zhao; Can Xu; Wei Wu; | |
280 | AttnIO: Knowledge Graph Exploration With In-and-Out Attention Flow For Knowledge-Grounded Dialogue Highlight: To this effect, we present AttnIO, a new dialog-conditioned path traversal model that makes a full use of rich structural information in KG based on two directions of attention flows. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Jaehun Jung; Bokyung Son; Sungwon Lyu; | |
281 | Amalgamating Knowledge From Two Teachers For Task-oriented Dialogue System With Adversarial Training Highlight: In this paper, we propose a Two-Teacher One-Student learning framework (TTOS) for task-oriented dialogue, with the goal of retrieving accurate KB entities and generating human-like responses simultaneously. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Wanwei He; Min Yang; Rui Yan; Chengming Li; Ying Shen; Ruifeng Xu; | |
282 | Task-oriented Domain-specific Meta-Embedding For Text Classification Highlight: In this paper, we propose a method to incorporate both domain-specific and task-oriented information into meta-embeddings. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Xin Wu; Yi Cai; Yang Kai; Tao Wang; Qing Li; | |
283 | Don’t Neglect The Obvious: On The Role Of Unambiguous Words In Word Sense Disambiguation Highlight: In this paper, we propose a simple method to provide annotations for most unambiguous words in a large corpus. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Daniel Loureiro; Jose Camacho-Collados; | |
284 | Within-Between Lexical Relation Classification Highlight: We propose the novel \textit{Within-Between} Relation model for recognizing lexical-semantic relations between words. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Oren Barkan; Avi Caciularu; Ido Dagan; | |
285 | With More Contexts Comes Better Performance: Contextualized Sense Embeddings For All-Round Word Sense Disambiguation Highlight: In this paper we present ARES (context-AwaRe Embeddings of Senses), a semi-supervised approach to producing sense embeddings for the lexical meanings within a lexical knowledge base that lie in a space that is comparable to that of contextualized word vectors. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Bianca Scarlini; Tommaso Pasini; Roberto Navigli; | code |
286 | Convolution Over Hierarchical Syntactic And Lexical Graphs For Aspect Level Sentiment Analysis Highlight: To tackle the above two limitations, we propose a novel architecture which convolutes over hierarchical syntactic and lexical graphs. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Mi Zhang; Tieyun Qian; | |
287 | Multi-Instance Multi-Label Learning Networks For Aspect-Category Sentiment Analysis Highlight: In this paper, we propose a Multi-Instance Multi-Label Learning Network for Aspect-Category sentiment analysis (AC-MIMLLN), which treats sentences as bags, words as instances, and the words indicating an aspect category as the key instances of the aspect category. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Yuncong Li; Cunxiang Yin; Sheng-hua Zhong; Xu Pan; | |
288 | Aspect Sentiment Classification With Aspect-Specific Opinion Spans Highlight: In this paper, we present a neat and effective structured attention model by aggregating multiple linear-chain CRFs. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Lu Xu; Lidong Bing; Wei Lu; Fei Huang; | |
289 | Emotion-Cause Pair Extraction As Sequence Labeling Based On A Novel Tagging Scheme Highlight: Targeting this issue, we regard the task as a sequence labeling problem and propose a novel tagging scheme with coding the distance between linked components into the tags, so that emotions and the corresponding causes can be extracted simultaneously. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Chaofa Yuan; Chuang Fan; Jianzhu Bao; Ruifeng Xu; | |
290 | End-to-End Emotion-Cause Pair Extraction Based On Sliding Window Multi-Label Learning Highlight: To tackle these shortcomings, we propose two joint frameworks for ECPE: 1) multi-label learning for the extraction of the cause clauses corresponding to the specified emotion clause (CMLL) and 2) multi-label learning for the extraction of the emotion clauses corresponding to the specified cause clause (EMLL). Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Zixiang Ding; Rui Xia; Jianfei Yu; | |
291 | Multi-modal Multi-label Emotion Detection With Modality And Label Dependence Highlight: In this paper, we focus on multi-label emotion detection in a multi-modal scenario. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Dong Zhang; Xincheng Ju; Junhui Li; Shoushan Li; Qiaoming Zhu; Guodong Zhou; | |
292 | Tasty Burgers, Soggy Fries: Probing Aspect Robustness In Aspect-Based Sentiment Analysis Highlight: To solve this problem, we develop a simple but effective approach to enrich ABSA test sets. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Xiaoyu Xing; Zhijing Jin; Di Jin; Bingning Wang; Qi Zhang; Xuanjing Huang; | code |
293 | Modeling Content Importance For Summarization With Pre-trained Language Models Highlight: In this work, we apply information theory on top of pre-trained language models and define the concept of importance from the perspective of information amount. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Liqiang Xiao; Lu Wang; Hao He; Yaohui Jin; | |
294 | Unsupervised Reference-Free Summary Quality Evaluation Via Contrastive Learning Highlight: In this work, we propose to evaluate the summary qualities without reference summaries by unsupervised contrastive learning. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Hanlu Wu; Tengfei Ma; Lingfei Wu; Tariro Manyumwa; Shouling Ji; | |
295 | Neural Extractive Summarization With Hierarchical Attentive Heterogeneous Graph Network Highlight: In this paper, we propose HAHSum (as shorthand for Hierarchical Attentive Heterogeneous Graph for Text Summarization), which well models different levels of information, including words and sentences, and spotlights redundancy dependencies between sentences. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Ruipeng Jia; Yanan Cao; Hengzhu Tang; Fang Fang; Cong Cao; Shi Wang; | |
296 | Coarse-to-Fine Query Focused Multi-Document Summarization Highlight: We propose a coarse-to-fine modeling framework which employs progressively more accurate modules for estimating whether text segments are relevant, likely to contain an answer, and central. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Yumo Xu; Mirella Lapata; | |
297 | Pre-training For Abstractive Document Summarization By Reinstating Source Text Highlight: This paper presents three sequence-to-sequence pre-training (in shorthand, STEP) objectives which allow us to pre-train a SEQ2SEQ based abstractive summarization model on unlabeled text. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Yanyan Zou; Xingxing Zhang; Wei Lu; Furu Wei; Ming Zhou; | |
298 | Learning From Context Or Names? An Empirical Study On Neural Relation Extraction Highlight: Based on the analyses, we propose an entity-masked contrastive pre-training framework for RE to gain a deeper understanding on both textual context and type information while avoiding rote memorization of entities or use of superficial cues in mentions. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Hao Peng; Tianyu Gao; Xu Han; Yankai Lin; Peng Li; Zhiyuan Liu; Maosong Sun; Jie Zhou; | code |
299 | SelfORE: Self-supervised Relational Feature Learning For Open Relation Extraction Highlight: In this work, we propose a self-supervised framework named SelfORE, which exploits weak, self-supervised signals by leveraging large pretrained language model for adaptive clustering on contextualized relational features, and bootstraps the self-supervised signals by improving contextualized features in relation classification. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Xuming Hu; Lijie Wen; Yusong Xu; Chenwei Zhang; Philip Yu; | |
300 | Denoising Relation Extraction From Document-level Distant Supervision Highlight: To alleviate this issue, we propose a novel pre-trained model for DocRE, which de-emphasize noisy DS data via multiple pre-training tasks. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Chaojun Xiao; Yuan Yao; Ruobing Xie; Xu Han; Zhiyuan Liu; Maosong Sun; Fen Lin; Leyu Lin; | |
301 | Let’s Stop Incorrect Comparisons In End-to-end Relation Extraction! Highlight: In this paper, we first identify several patterns of invalid comparisons in published papers and describe them to avoid their propagation. We then propose a small empirical study to quantify the most common mistake’s impact and evaluate it leads to overestimating the final RE performance by around 5% on ACE05. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Bruno Taillé; Vincent Guigue; Geoffrey Scoutheeten; Patrick Gallinari; | |
302 | Exposing Shallow Heuristics Of Relation Extraction Models With Challenge Data Highlight: We identify failure modes of SOTA relation extraction (RE) models trained on TACRED, which we attribute to limitations in the data annotation process. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Shachar Rosenman; Alon Jacovi; Yoav Goldberg; | |
303 | Global-to-Local Neural Networks For Document-Level Relation Extraction Highlight: In this paper, we propose a novel model to document-level RE, by encoding the document information in terms of entity global and local representations as well as context relation representations. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Difeng Wang; Wei Hu; Ermei Cao; Weijian Sun; | |
304 | Recurrent Interaction Network For Jointly Extracting Entities And Classifying Relations Highlight: As a solution, we design a multi-task learning model which we refer to as recurrent interaction network which allows the learning of interactions dynamically, to effectively model task-specific features for classification. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Kai Sun; Richong Zhang; Samuel Mensah; Yongyi Mao; Xudong Liu; | |
305 | Temporal Knowledge Base Completion: New Algorithms And Evaluation Protocols Highlight: In response, we propose improved TKBC evaluation protocols for both link and time prediction tasks, dealing with subtle issues that arise from the partial overlap of time intervals in gold instances and system predictions. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Prachi Jain; Sushant Rathi; Mausam; Soumen Chakrabarti; | |
306 | OpenIE6: Iterative Grid Labeling And Coordination Analysis For Open Information Extraction Highlight: In this paper, we bridge this trade-off by presenting an iterative labeling-based system that establishes a new state of the art for OpenIE, while extracting 10x faster. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Keshav Kolluru; Vaibhav Adlakha; Samarth Aggarwal; Mausam; Soumen Chakrabarti; | |
307 | Public Sentiment Drift Analysis Based On Hierarchical Variational Auto-encoder Highlight: In this paper, we focus on distribution learning by proposing a novel Hierarchical Variational Auto-Encoder (HVAE) model to learn better distribution representation, and design a new drift measure to directly evaluate distribution changes between historical data and new data. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Wenyue Zhang; Xiaoli Li; Yang Li; Suge Wang; Deyu Li; Jian Liao; Jianxing Zheng; | |
308 | Point To The Expression: Solving Algebraic Word Problems Using The Expression-Pointer Transformer Model Highlight: To address each of these two issues, we propose a pure neural model, Expression-Pointer Transformer (EPT), which uses (1) Expression’ token and (2) operand-context pointers when generating solution equations. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Bugeun Kim; Kyung Seo Ki; Donggeon Lee; Gahgene Gweon; | |
309 | Semantically-Aligned Universal Tree-Structured Solver For Math Word Problems Highlight: Herein, we propose a simple but efficient method called Universal Expression Tree (UET) to make the first attempt to represent the equations of various MWPs uniformly. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Jinghui Qin; Lihui Lin; Xiaodan Liang; Rumin Zhang; Liang Lin; | |
310 | Neural Topic Modeling By Incorporating Document Relationship Graph Highlight: In this paper, we propose Graph Topic Model (GTM), a GNN based neural topic model that represents a corpus as a document relationship graph. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Deyu Zhou; Xuemeng Hu; Rui Wang; | |
311 | Routing Enforced Generative Model For Recipe Generation Highlight: In this work, we propose a routing method to dive into the content selection under the internal restrictions. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Zhiwei Yu; Hongyu Zang; Xiaojun Wan; | |
312 | Assessing The Helpfulness Of Learning Materials With Inference-Based Learner-Like Agent Highlight: Thus, we propose the inference-based learner-like agent to mimic learner behavior and identify good learning materials by examining the agent’s performance. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Yun-Hsuan Jen; Chieh-Yang Huang; MeiHua Chen; Ting-Hao Huang; Lun-Wei Ku; | |
313 | Selection And Generation: Learning Towards Multi-Product Advertisement Post Generation Highlight: We propose a novel end-to-end model named S-MG Net to generate the AD post. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Zhangming Chan; Yuchi Zhang; Xiuying Chen; Shen Gao; Zhiqiang Zhang; Dongyan Zhao; Rui Yan; | |
314 | Form2Seq : A Framework For Higher-Order Form Structure Extraction Highlight: To mitigate this, we propose Form2Seq, a novel sequence-to-sequence (Seq2Seq) inspired framework for structure extraction using text, with a specific focus on forms, which leverages relative spatial arrangement of structures. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Milan Aggarwal; Hiresh Gupta; Mausoom Sarkar; Balaji Krishnamurthy; | |
315 | Domain Adaptation Of Thai Word Segmentation Models Using Stacked Ensemble Highlight: We propose a filter-and-refine solution based on the stacked-ensemble learning paradigm to address this black-box limitation. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Peerat Limkonchotiwat; Wannaphong Phatthiyaphaibun; Raheem Sarwar; Ekapol Chuangsuwanich; Sarana Nutanong; | |
316 | DagoBERT: Generating Derivational Morphology With A Pretrained Language Model Highlight: We present the first study investigating this question, taking BERT as the example PLM. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Valentin Hofmann; Janet Pierrehumbert; Hinrich Schütze; | |
317 | Attention Is All You Need For Chinese Word Segmentation Highlight: Taking greedy decoding algorithm as it should be, this work focuses on further strengthening the model itself for Chinese word segmentation (CWS), which results in an even more fast and more accurate CWS model. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Sufeng Duan; Hai Zhao; | |
318 | A Joint Multiple Criteria Model In Transfer Learning For Cross-domain Chinese Word Segmentation Highlight: To this end, we propose a joint multiple criteria model that shares all parameters to integrate different segmentation criteria into one model. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Kaiyu Huang; Degen Huang; Zhuang Liu; Fengran Mo; | |
319 | Alignment-free Cross-lingual Semantic Role Labeling Highlight: We propose a cross-lingual SRL model which only requires annotations in a source language and access to raw text in the form of a parallel corpus. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Rui Cai; Mirella Lapata; | |
320 | Leveraging Declarative Knowledge In Text And First-Order Logic For Fine-Grained Propaganda Detection Highlight: Instead of merely learning from input-output datapoints in training data, we introduce an approach to inject declarative knowledge of fine-grained propaganda techniques. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Ruize Wang; Duyu Tang; Nan Duan; Wanjun Zhong; Zhongyu Wei; Xuanjing Huang; Daxin Jiang; Ming Zhou; | |
321 | X-SRL: A Parallel Cross-Lingual Semantic Role Labeling Dataset Highlight: In this work we propose a method to automatically construct an SRL corpus that is parallel in four languages: English, French, German, Spanish, with unified predicate and role annotations that are fully comparable across languages. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Angel Daza; Anette Frank; | |
322 | Graph Convolutions Over Constituent Trees For Syntax-Aware Semantic Role Labeling Highlight: In contrast, we show how graph convolutional networks (GCNs) can be used to encode constituent structures and inform an SRL system. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Diego Marcheggiani; Ivan Titov; | |
323 | Fast Semantic Parsing With Well-typedness Guarantees Highlight: We describe an A* parser and a transition-based parser for AM dependency parsing which guarantee well-typedness and improve parsing speed by up to 3 orders of magnitude, while maintaining or improving accuracy. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Matthias Lindemann; Jonas Groschwitz; Alexander Koller; | |
324 | Improving Out-of-Scope Detection In Intent Classification By Using Embeddings Of The Word Graph Space Of The Classes Highlight: This paper explores how intent classification can be improved by representing the class labels not as a discrete set of symbols but as a space where the word graphs associated to each class are mapped using typical graph embedding techniques. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Paulo Cavalin; Victor Henrique Alves Ribeiro; Ana Appel; Claudio Pinhanez; | |
325 | Supervised Seeded Iterated Learning For Interactive Language Learning Highlight: Given these observations, we introduce Supervised Seeded Iterated Learning (SSIL) to combine both methods to minimize their respective weaknesses. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Yuchen Lu; Soumye Singhal; Florian Strub; Olivier Pietquin; Aaron Courville; | |
326 | Spot The Bot: A Robust And Efficient Framework For The Evaluation Of Conversational Dialogue Systems Highlight: In this work, we introduce Spot The Bot, a cost-efficient and robust evaluation framework that replaces human-bot conversations with conversations between bots. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Jan Deriu; Don Tuggener; Pius von Däniken; Jon Ander Campos; Alvaro Rodrigo; Thiziri Belkacem; Aitor Soroa; Eneko Agirre; Mark Cieliebak; | |
327 | Human-centric Dialog Training Via Offline Reinforcement Learning Highlight: We solve the challenge by developing a novel class of offline RL algorithms. These algorithms use KL-control to penalize divergence from a pre-trained prior language model, and use a new strategy to make the algorithm pessimistic, instead of optimistic, in the face of uncertainty. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Natasha Jaques; Judy Hanwen Shen; Asma Ghandeharioun; Craig Ferguson; Agata Lapedriza; Noah Jones; Shixiang Gu; Rosalind Picard; | |
328 | Speakers Fill Lexical Semantic Gaps With Context Highlight: To investigate whether this is the case, we operationalise the lexical ambiguity of a word as the entropy of meanings it can take, and provide two ways to estimate this-one which requires human annotation (using WordNet), and one which does not (using BERT), making it readily applicable to a large number of languages. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Tiago Pimentel; Rowan Hall Maudslay; Damian Blasi; Ryan Cotterell; | |
329 | Investigating Cross-Linguistic Adjective Ordering Tendencies With A Latent-Variable Model Highlight: We present the first purely corpus-driven model of multi-lingual adjective ordering in the form of a latent-variable model that can accurately order adjectives across 24 different languages, even when the training and testing languages are different. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Jun Yen Leung; Guy Emerson; Ryan Cotterell; | |
330 | Surprisal Predicts Code-Switching In Chinese-English Bilingual Text Highlight: We describe and model a new dataset of Chinese-English text with 1476 clean code-switched sentences, translated back into Chinese. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Jesús Calvillo; Le Fang; Jeremy Cole; David Reitter; | |
331 | Word Frequency Does Not Predict Grammatical Knowledge In Language Models Highlight: In this work, we investigate whether there are systematic sources of variation in the language models’ accuracy. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Charles Yu; Ryan Sie; Nicolas Tedeschi; Leon Bergen; | |
332 | Improving Word Sense Disambiguation With Translations Highlight: In this paper, we present a novel approach that improves the performance of a base WSD system using machine translation. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Yixing Luan; Bradley Hauer; Lili Mou; Grzegorz Kondrak; | |
333 | Towards Better Context-aware Lexical Semantics:Adjusting Contextualized Representations Through Static Anchors Highlight: In this paper, we present a post-processing technique that enhances these representations by learning a transformation through static anchors. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Qianchu Liu; Diana McCarthy; Anna Korhonen; | |
334 | Compositional Demographic Word Embeddings Highlight: We propose a new form of personalized word embeddings that use demographic-specific word representations derived compositionally from full or partial demographic information for a user (i.e., gender, age, location, religion). Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Charles Welch; Jonathan K. Kummerfeld; Verónica Pérez-Rosas; Rada Mihalcea; | |
335 | Do “Undocumented Workers” == “Illegal Aliens”? Differentiating Denotation And Connotation In Vector Spaces Highlight: In this study, we propose an adversarial nerual netowrk that decomposes a pretrained representation as independent denotation and connotation representations. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Albert Webson; Zhizhong Chen; Carsten Eickhoff; Ellie Pavlick; | |
336 | Multi-View Sequence-to-Sequence Models With Conversational Structure For Abstractive Dialogue Summarization Highlight: This work proposes a multi-view sequence-to-sequence model by first extracting conversational structures of unstructured daily chats from different views to represent conversations and then utilizing a multi-view decoder to incorporate different views to generate dialogue summaries. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Jiaao Chen; Diyi Yang; | code |
337 | Few-Shot Learning For Opinion Summarization Highlight: In this work, we show that even a handful of summaries is sufficient to bootstrap generation of the summary text with all expected properties, such as writing style, informativeness, fluency, and sentiment preservation. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Arthur Bražinskas; Mirella Lapata; Ivan Titov; | |
338 | Learning To Fuse Sentences With Transformers For Summarization Highlight: In this paper, we explore the ability of Transformers to fuse sentences and propose novel algorithms to enhance their ability to perform sentence fusion by leveraging the knowledge of points of correspondence between sentences. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Logan Lebanoff; Franck Dernoncourt; Doo Soon Kim; Lidan Wang; Walter Chang; Fei Liu; | |
339 | Stepwise Extractive Summarization And Planning With Structured Transformers Highlight: We propose encoder-centric stepwise models for extractive summarization using structured transformers – HiBERT and Extended Transformers. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Shashi Narayan; Joshua Maynez; Jakub Adamek; Daniele Pighin; Blaz Bratanic; Ryan McDonald; | |
340 | CLIRMatrix: A Massively Large Collection Of Bilingual And Multilingual Datasets For Cross-Lingual Information Retrieval Highlight: We present CLIRMatrix, a massively large collection of bilingual and multilingual datasets for Cross-Lingual Information Retrieval extracted automatically from Wikipedia. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Shuo Sun; Kevin Duh; | |
341 | SLEDGE-Z: A Zero-Shot Baseline For COVID-19 Literature Search Highlight: In this work, we present a zero-shot ranking algorithm that adapts to COVID-related scientific literature. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Sean MacAvaney; Arman Cohan; Nazli Goharian; | |
342 | Modularized Transfomer-based Ranking Framework Highlight: In this work, we modularize the Transformer ranker into separate modules for text representation and interaction. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Luyu Gao; Zhuyun Dai; Jamie Callan; | |
343 | Ad-hoc Document Retrieval Using Weak-Supervision With BERT And GPT2 Highlight: We describe a weakly-supervised method for training deep learning models for the task of ad-hoc document retrieval. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Yosi Mass; Haggai Roitman; | |
344 | Adversarial Semantic Collisions Highlight: We develop gradient-based approaches for generating semantic collisions and demonstrate that state-of-the-art models for many tasks which rely on analyzing the meaning and similarity of texts-including paraphrase identification, document retrieval, response suggestion, and extractive summarization-are vulnerable to semantic collisions. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Congzheng Song; Alexander Rush; Vitaly Shmatikov; | code |
345 | Learning Explainable Linguistic Expressions With Neural Inductive Logic Programming For Sentence Classification Highlight: We present RuleNN, a neural network architecture for learning transparent models for sentence classification. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Prithviraj Sen; Marina Danilevsky; Yunyao Li; Siddhartha Brahma; Matthias Boehm; Laura Chiticariu; Rajasekar Krishnamurthy; | |
346 | AutoPrompt: Eliciting Knowledge From Language Models With Automatically Generated Prompts Highlight: To address this, we develop AutoPrompt, an automated method to create prompts for a diverse set of tasks, based on a gradient-guided search. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Taylor Shin; Yasaman Razeghi; Robert L. Logan IV; Eric Wallace; Sameer Singh; | |
347 | Learning Variational Word Masks To Improve The Interpretability Of Neural Text Classifiers Highlight: To address this limitation, we propose the variational word mask (VMASK) method to automatically learn task-specific important words and reduce irrelevant information on classification, which ultimately improves the interpretability of model predictions. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Hanjie Chen; Yangfeng Ji; | |
348 | Sparse Text Generation Highlight: In this paper, we use the recently introduced entmax transformation to train and sample from a natively sparse language model, avoiding this mismatch. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Pedro Henrique Martins; Zita Marinho; André F. T. Martins; | |
349 | PlotMachines: Outline-Conditioned Generation With Dynamic Plot State Tracking Highlight: We present PlotMachines, a neural narrative model that learns to transform an outline into a coherent story by tracking the dynamic plot states. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Hannah Rashkin; Asli Celikyilmaz; Yejin Choi; Jianfeng Gao; | |
350 | Do Sequence-to-sequence VAEs Learn Global Features Of Sentences? Highlight: To alleviate this, we investigate alternative architectures based on bag-of-words assumptions and language model pretraining. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Tom Bosc; Pascal Vincent; | |
351 | Content Planning For Neural Story Generation With Aristotelian Rescoring Highlight: We utilize a plot-generation language model along with an ensemble of rescoring models that each implement an aspect of good story-writing as detailed in Aristotle’s Poetics. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Seraphina Goldfarb-Tarrant; Tuhin Chakrabarty; Ralph Weischedel; Nanyun Peng; | |
352 | Generating Dialogue Responses From A Semantic Latent Space Highlight: In this work, we hypothesize that the current models are unable to integrate information from multiple semantically similar valid responses of a prompt, resulting in the generation of generic and uninformative responses. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Wei-Jen Ko; Avik Ray; Yilin Shen; Hongxia Jin; | |
353 | Refer, Reuse, Reduce: Generating Subsequent References In Visual And Conversational Contexts Highlight: We propose a generation model that produces referring utterances grounded in both the visual and the conversational context. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Ece Takmaz; Mario Giulianelli; Sandro Pezzelle; Arabella Sinclair; Raquel Fernández; | |
354 | Visually Grounded Compound PCFGs Highlight: In this work, we study visually grounded grammar induction and learn a constituency parser from both unlabeled text and its visual groundings. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Yanpeng Zhao; Ivan Titov; | |
355 | ALICE: Active Learning With Contrastive Natural Language Explanations Highlight: We propose Active Learning with Contrastive Explanations (ALICE), an expert-in-the-loop training framework that utilizes contrastive natural language explanations to improve data efficiency in learning. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Weixin Liang; James Zou; Zhou Yu; | |
356 | Room-Across-Room: Multilingual Vision-and-Language Navigation With Dense Spatiotemporal Grounding Highlight: We introduce Room-Across-Room (RxR), a new Vision-and-Language Navigation (VLN) dataset. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Alexander Ku; Peter Anderson; Roma Patel; Eugene Ie; Jason Baldridge; | |
357 | SSCR: Iterative Language-Based Image Editing Via Self-Supervised Counterfactual Reasoning Highlight: In this paper, we introduce a Self-Supervised Counterfactual Reasoning (SSCR) framework that incorporates counterfactual thinking to overcome data scarcity. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Tsu-Jui Fu; Xin Wang; Scott Grafton; Miguel Eckstein; William Yang Wang; | |
358 | Identifying Elements Essential For BERT’s Multilinguality Highlight: We aim to identify architectural properties of BERT and linguistic properties of languages that are necessary for BERT to become multilingual. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Philipp Dufter; Hinrich Schütze; | |
359 | On Negative Interference In Multilingual Models: Findings And A Meta-Learning Treatment Highlight: In this paper, we present the first systematic study of negative interference. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Zirui Wang; Zachary C. Lipton; Yulia Tsvetkov; | |
360 | Pre-tokenization Of Multi-word Expressions In Cross-lingual Word Embeddings Highlight: We propose a simple method for word translation of MWEs to and from English in ten languages: we first compile lists of MWEs in each language and then tokenize the MWEs as single tokens before training word embeddings. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Naoki Otani; Satoru Ozaki; Xingyuan Zhao; Yucen Li; Micaelah St Johns; Lori Levin; | |
361 | Monolingual Adapters For Zero-Shot Neural Machine Translation Highlight: We propose a novel adapter layer formalism for adapting multilingual models. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Jerin Philip; Alexandre Berard; Matthias Gallé; Laurent Besacier; | |
362 | Do Explicit Alignments Robustly Improve Multilingual Encoders? Highlight: In this paper, we propose a new contrastive alignment objective that can better utilize such signal, and examine whether these previous alignment methods can be adapted to noisier sources of aligned data: a randomly sampled 1 million pair subset of the OPUS collection. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Shijie Wu; Mark Dredze; | |
363 | From Zero To Hero: On The Limitations Of Zero-Shot Language Transfer With Multilingual Transformers Highlight: In this work, we analyze the limitations of downstream language transfer with MMTs, showing that, much like cross-lingual word embeddings, they are substantially less effective in resource-lean scenarios and for distant languages. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Anne Lauscher; Vinit Ravishankar; Ivan Vulić; Goran Glavaš; | |
364 | Distilling Multiple Domains For Neural Machine Translation Highlight: In this paper, we propose a framework for training a single multi-domain neural machine translation model that is able to translate several domains without increasing inference time or memory usage. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Anna Currey; Prashant Mathur; Georgiana Dinu; | |
365 | Making Monolingual Sentence Embeddings Multilingual Using Knowledge Distillation Highlight: We present an easy and efficient method to extend existing sentence embedding models to new languages. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Nils Reimers; Iryna Gurevych; | |
366 | A Streaming Approach For Efficient Batched Beam Search Highlight: We propose an efficient batching strategy for variable-length decoding on GPU architectures. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Kevin Yang; Violet Yao; John DeNero; Dan Klein; | |
367 | Improving Multilingual Models With Language-Clustered Vocabularies Highlight: In this work, we introduce a novel procedure for multilingual vocabulary generation that combines the separately trained vocabularies of several automatically derived language clusters, thus balancing the trade-off between cross-lingual subword sharing and language-specific vocabularies. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Hyung Won Chung; Dan Garrette; Kiat Chuan Tan; Jason Riesa; | |
368 | Zero-Shot Cross-Lingual Transfer With Meta Learning Highlight: We show that this challenging setup can be approached using meta-learning: in addition to training a source language model, another model learns to select which training instances are the most beneficial to the first. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Farhad Nooralahzadeh; Giannis Bekoulis; Johannes Bjerva; Isabelle Augenstein; | code |
369 | The Multilingual Amazon Reviews Corpus Highlight: We present the Multilingual Amazon Reviews Corpus (MARC), a large-scale collection of Amazon reviews for multilingual text classification. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Phillip Keung; Yichao Lu; György Szarvas; Noah A. Smith; | |
370 | GLUCOSE: GeneraLized And COntextualized Story Explanations Highlight: As a step toward AI systems that can build similar mental models, we introduce GLUCOSE, a large-scale dataset of implicit commonsense causal knowledge, encoded as causal mini-theories about the world, each grounded in a narrative context. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Nasrin Mostafazadeh; Aditya Kalyanpur; Lori Moon; David Buchanan; Lauren Berkowitz; Or Biran; Jennifer Chu-Carroll; | |
371 | Character-level Representations Improve DRS-based Semantic Parsing Even In The Age Of BERT Highlight: We combine character-level and contextual language model representations to improve performance on Discourse Representation Structure parsing. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Rik van Noord; Antonio Toral; Johan Bos; | |
372 | Infusing Disease Knowledge Into BERT For Health Question Answering, Medical Inference And Disease Name Recognition Highlight: Specifically, we propose a new disease knowledge infusion training procedure and evaluate it on a suite of BERT models including BERT, BioBERT, SciBERT, ClinicalBERT, BlueBERT, and ALBERT. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Yun He; Ziwei Zhu; Yin Zhang; Qin Chen; James Caverlee; | |
373 | Unsupervised Commonsense Question Answering With Self-Talk Highlight: We propose an unsupervised framework based on self-talk as a novel alternative to multiple-choice commonsense tasks. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Vered Shwartz; Peter West; Ronan Le Bras; Chandra Bhagavatula; Yejin Choi; | |
374 | Reasoning About Goals, Steps, And Temporal Ordering With WikiHow Highlight: We propose a suite of reasoning tasks on two types of relations between procedural events: goal-step relations (learn poses is a step in the larger goal of doing yoga) and step-step temporal relations (buy a yoga mat typically precedes learn poses). Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Li Zhang; Qing Lyu; Chris Callison-Burch; | |
375 | Structural Supervision Improves Few-Shot Learning And Syntactic Generalization In Neural Language Models Highlight: We find that in most cases, the neural models are able to induce the proper syntactic generalizations after minimal exposure, often from just two examples during training, and that the two structurally supervised models generalize more accurately than the LSTM model. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Ethan Wilcox; Peng Qian; Richard Futrell; Ryosuke Kohita; Roger Levy; Miguel Ballesteros; | |
376 | Investigating Representations Of Verb Bias In Neural Language Models Highlight: Here we introduce DAIS, a large benchmark dataset containing 50K human judgments for 5K distinct sentence pairs in the English dative alternation. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Robert Hawkins; Takateru Yamakoshi; Thomas Griffiths; Adele Goldberg; | |
377 | Generating Image Descriptions Via Sequential Cross-Modal Alignment Guided By Human Gaze Highlight: In this paper, we investigate such sequential cross-modal alignment by modelling the image description generation process computationally. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Ece Takmaz; Sandro Pezzelle; Lisa Beinborn; Raquel Fernández; | |
378 | Optimus: Organizing Sentences Via Pre-trained Modeling Of A Latent Space Highlight: In this paper, we propose the first large-scale language VAE model Optimus (Organizing sentences via Pre-Trained Modeling of a Universal Space). Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Chunyuan Li; Xiang Gao; Yuan Li; Baolin Peng; Xiujun Li; Yizhe Zhang; Jianfeng Gao; | |
379 | BioMegatron: Larger Biomedical Domain Language Model Highlight: We empirically study and evaluate several factors that can affect performance on domain language applications, such as the sub-word vocabulary set, model size, pre-training corpus, and domain transfer. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Hoo-Chang Shin; Yang Zhang; Evelina Bakhturina; Raul Puri; Mostofa Patwary; Mohammad Shoeybi; Raghav Mani; | |
380 | Text Segmentation By Cross Segment Attention Highlight: In this work, we propose three transformer-based architectures and provide comprehensive comparisons with previously proposed approaches on three standard datasets. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Michal Lukasik; Boris Dadachev; Kishore Papineni; Gonçalo Simões; | |
381 | RussianSuperGLUE: A Russian Language Understanding Evaluation Benchmark Highlight: In this paper, we introduce an advanced Russian general language understanding evaluation benchmark – Russian SuperGLUE. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Tatiana Shavrina; Alena Fenogenova; Emelyanov Anton; Denis Shevelev; Ekaterina Artemova; Valentin Malykh; Vladislav Mikhailov; Maria Tikhonova; Andrey Chertok; Andrey Evlampiev; | |
382 | An Empirical Study Of Pre-trained Transformers For Arabic Information Extraction Highlight: In this paper, we pre-train a customized bilingual BERT, dubbed GigaBERT, that is designed specifically for Arabic NLP and English-to-Arabic zero-shot transfer learning. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Wuwei Lan; Yang Chen; Wei Xu; Alan Ritter; | code |
383 | TNT: Text Normalization Based Pre-training Of Transformers For Content Moderation Highlight: In this work, we present a new language pre-training model TNT (Text Normalization based pre-training of Transformers) for content moderation. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Fei Tan; Yifan Hu; Changwei Hu; Keqian Li; Kevin Yen; | |
384 | Methods For Numeracy-Preserving Word Embeddings Highlight: We propose a new methodology to assign and learn embeddings for numbers. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Dhanasekar Sundararaman; Shijing Si; Vivek Subramanian; Guoyin Wang; Devamanyu Hazarika; Lawrence Carin; | |
385 | An Empirical Investigation Of Contextualized Number Prediction Highlight: Specifically, we introduce a suite of output distribution parameterizations that incorporate latent variables to add expressivity and better fit the natural distribution of numeric values in running text, and combine them with both recur-rent and transformer-based encoder architectures. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Taylor Berg-Kirkpatrick; Daniel Spokoyny; | |
386 | Modeling The Music Genre Perception Across Language-Bound Cultures Highlight: In this work, we study the feasibility of obtaining relevant cross-lingual, culture-specific music genre annotations based only on language-specific semantic representations, namely distributed concept embeddings and ontologies. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Elena V. Epure; Guillaume Salha; Manuel Moussallam; Romain Hennequin; | |
387 | Joint Estimation And Analysis Of Risk Behavior Ratings In Movie Scripts Highlight: To address this limitation, we propose a model that estimates content ratings based on the language use in movie scripts, making our solution available at the earlier stages of creative production. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Victor Martinez; Krishna Somandepalli; Yalda Tehranian-Uhls; Shrikanth Narayanan; | |
388 | Keep It Surprisingly Simple: A Simple First Order Graph Based Parsing Model For Joint Morphosyntactic Parsing In Sanskrit Highlight: We propose a graph-based model for joint morphological parsing and dependency parsing in Sanskrit. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Amrith Krishna; Ashim Gupta; Deepak Garasangi; Pavankumar Satuluri; Pawan Goyal; | |
389 | Unsupervised Parsing Via Constituency Tests Highlight: We propose a method for unsupervised parsing based on the linguistic notion of a constituency test. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Steven Cao; Nikita Kitaev; Dan Klein; | |
390 | Please Mind The Root: Decoding Arborescences For Dependency Parsing Highlight: We analyzed the output of state-of-the-art parsers on many languages from the Universal Dependency Treebank: although these parsers are often able to learn that trees which violate the constraint should be assigned lower probabilities, their ability to do so unsurprisingly de-grades as the size of the training set decreases. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Ran Zmigrod; Tim Vieira; Ryan Cotterell; | |
391 | Unsupervised Cross-Lingual Part-of-Speech Tagging For Truly Low-Resource Scenarios Highlight: We describe a fully unsupervised cross-lingual transfer approach for part-of-speech (POS) tagging under a truly low resource scenario. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Ramy Eskander; Smaranda Muresan; Michael Collins; | |
392 | Unsupervised Parsing With S-DIORA: Single Tree Encoding For Deep Inside-Outside Recursive Autoencoders Highlight: In this paper, we discover that while DIORA exhaustively encodes all possible binary trees of a sentence with a soft dynamic program, its vector averaging approach is locally greedy and cannot recover from errors when computing the highest scoring parse tree in bottom-up chart parsing. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Andrew Drozdov; Subendhu Rongali; Yi-Pei Chen; Tim O’Gorman; Mohit Iyyer; Andrew McCallum; | |
393 | Utility Is In The Eye Of The User: A Critique Of NLP Leaderboards Highlight: In this opinion paper, we study the divergence between what is incentivized by leaderboards and what is useful in practice through the lens of microeconomic theory. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Kawin Ethayarajh; Dan Jurafsky; | |
394 | An Empirical Investigation Towards Efficient Multi-Domain Language Model Pre-training Highlight: In this paper we conduct an empirical investigation into known methods to mitigate CF. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Kristjan Arumae; Qing Sun; Parminder Bhatia; | |
395 | Analyzing Individual Neurons In Pre-trained Language Models Highlight: We found small subsets of neurons to predict linguistic tasks, with lower level tasks (such as morphology) localized in fewer neurons, compared to higher level task of predicting syntax. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Nadir Durrani; Hassan Sajjad; Fahim Dalvi; Yonatan Belinkov; | |
396 | Dissecting Span Identification Tasks With Performance Prediction Highlight: Our contributions are: (a) we identify key properties of span ID tasks that can inform performance prediction; (b) we carry out a large-scale experiment on English data, building a model to predict performance for unseen span ID tasks that can support architecture choices; (c), we investigate the parameters of the meta model, yielding new insights on how model and task properties interact to affect span ID performance. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Sean Papay; Roman Klinger; Sebastian Padó; | |
397 | Assessing Phrasal Representation And Composition In Transformers Highlight: In this paper, we present systematic analysis of phrasal representations in state-of-the-art pre-trained transformers. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Lang Yu; Allyson Ettinger; | |
398 | Analyzing Redundancy In Pretrained Transformer Models Highlight: In this paper, we study the cause of these limitations by defining a notion of Redundancy, which we categorize into two classes: General Redundancy and Task-specific Redundancy. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Fahim Dalvi; Hassan Sajjad; Nadir Durrani; Yonatan Belinkov; | |
399 | Be More With Less: Hypergraph Attention Networks For Inductive Text Classification Highlight: To address those issues, in this paper, we propose a principled model – hypergraph attention networks (HyperGAT), which can obtain more expressive power with less computational consumption for text representation learning. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Kaize Ding; Jianling Wang; Jundong Li; Dingcheng Li; Huan Liu; | |
400 | Entities As Experts: Sparse Memory Access With Entity Supervision Highlight: We introduce a new model-Entities as Experts (EaE)-that can access distinct memories of the entities mentioned in a piece of text. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Thibault Févry; Livio Baldini Soares; Nicholas FitzGerald; Eunsol Choi; Tom Kwiatkowski; | |
401 | H2KGAT: Hierarchical Hyperbolic Knowledge Graph Attention Network Highlight: To fill this gap, in this paper, we propose Hierarchical Hyperbolic Knowledge Graph Attention Network (H2KGAT), a novel knowledge graph embedding framework, which is able to better model and infer hierarchical relation patterns. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Shen Wang; Xiaokai Wei; Cicero Nogueira dos Santos; Zhiguo Wang; Ramesh Nallapati; Andrew Arnold; Bing Xiang; Philip S. Yu; | |
402 | Does The Objective Matter? Comparing Training Objectives For Pronoun Resolution Highlight: In this work, we make a fair comparison of the performance and seed-wise stability of four models that represent the four categories of objectives. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Yordan Yordanov; Oana-Maria Camburu; Vid Kocijan; Thomas Lukasiewicz; | |
403 | On Losses For Modern Language Models Highlight: In this paper, we 1) clarify NSP’s effect on BERT pre-training, 2) explore fourteen possible auxiliary pre-training tasks, of which seven are novel to modern language models, and 3) investigate different ways to include multiple tasks into pre-training. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Stéphane Aroca-Ouellette; Frank Rudzicz; | |
404 | We Can Detect Your Bias: Predicting The Political Ideology Of News Articles Highlight: From a modeling perspective, we propose an adversarial media adaptation, as well as a specially adapted triplet loss. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Ramy Baly; Giovanni Da San Martino; James Glass; Preslav Nakov; | |
405 | Semantic Label Smoothing For Sequence To Sequence Problems Highlight: Unlike these works, in this paper, we propose a technique that smooths over \textit{well formed} relevant sequences that not only have sufficient n-gram overlap with the target sequence, but are also \textit{semantically similar}. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Michal Lukasik; Himanshu Jain; Aditya Menon; Seungyeon Kim; Srinadh Bhojanapalli; Felix Yu; Sanjiv Kumar; | |
406 | Training For Gibbs Sampling On Conditional Random Fields With Neural Scoring Factors Highlight: In this work, we present an approach for efficiently training and decoding hybrids of graphical models and neural networks based on Gibbs sampling. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Sida Gao; Matthew R. Gormley; | |
407 | Multilevel Text Alignment With Cross-Document Attention Highlight: We propose a new learning approach that equips previously established hierarchical attention encoders for representing documents with a cross-document attention component, enabling structural comparisons across different levels (document-to-document and sentence-to-document). Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Xuhui Zhou; Nikolaos Pappas; Noah A. Smith; | |
408 | Conversational Semantic Parsing Highlight: In this paper, we propose a semantic representation for such task-oriented conversational systems that can represent concepts such as co-reference and context carryover, enabling comprehensive understanding of queries in a session. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Armen Aghajanyan; Jean Maillard; Akshat Shrivastava; Keith Diedrick; Michael Haeger; Haoran Li; Yashar Mehdad; Veselin Stoyanov; Anuj Kumar; Mike Lewis; Sonal Gupta; | |
409 | Probing Task-Oriented Dialogue Representation From Language Models Highlight: The goals of this empirical paper are to 1) investigate probing techniques, especially from the unsupervised mutual information aspect, 2) provide guidelines of pre-trained language model selection for the dialogue research community, 3) find insights of pre-training factors for dialogue application that may be the key to success. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Chien-Sheng Wu; Caiming Xiong; | |
410 | End-to-End Slot Alignment And Recognition For Cross-Lingual NLU Highlight: In this work, we propose a novel end-to-end model that learns to align and predict target slot labels jointly for cross-lingual transfer. We introduce MultiATIS++, a new multilingual NLU corpus that extends the Multilingual ATIS corpus to nine languages across four language families, and evaluate our method using the corpus. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Weijia Xu; Batool Haider; Saab Mansour; | |
411 | Discriminative Nearest Neighbor Few-Shot Intent Detection By Transferring Natural Language Inference Highlight: In this paper, we present a simple yet effective approach, discriminative nearest neighbor classification with deep self-attention. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Jianguo Zhang; Kazuma Hashimoto; Wenhao Liu; Chien-Sheng Wu; Yao Wan; Philip Yu; Richard Socher; Caiming Xiong; | code |
412 | Simple Data Augmentation With The Mask Token Improves Domain Adaptation For Dialog Act Tagging Highlight: In this work, we investigate how to better adapt DA taggers to desired target domains with only unlabeled data. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Semih Yavuz; Kazuma Hashimoto; Wenhao Liu; Nitish Shirish Keskar; Richard Socher; Caiming Xiong; | |
413 | Low-Resource Domain Adaptation For Compositional Task-Oriented Semantic Parsing Highlight: In this paper, we focus on adapting task-oriented semantic parsers to low-resource domains, and propose a novel method that outperforms a supervised neural model at a 10-fold data reduction. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Xilun Chen; Asish Ghoshal; Yashar Mehdad; Luke Zettlemoyer; Sonal Gupta; | |
414 | Sound Natural: Content Rephrasing In Dialog Systems Highlight: In this paper, we study the problem of rephrasing with messaging as a use case and release a dataset of 3000 pairs of original query and rephrased query. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Arash Einolghozati; Anchit Gupta; Keith Diedrick; Sonal Gupta; | |
415 | Zero-Shot Crosslingual Sentence Simplification Highlight: We propose a zero-shot modeling framework which transfers simplification knowledge from English to another language (for which no parallel simplification corpus exists) while generalizing across languages and tasks. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Jonathan Mallinson; Rico Sennrich; Mirella Lapata; | |
416 | Facilitating The Communication Of Politeness Through Fine-Grained Paraphrasing Highlight: In this work, we take the first steps towards automatically assisting people in adjusting their language to a specific communication circumstance. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Liye Fu; Susan Fussell; Cristian Danescu-Niculescu-Mizil; | |
417 | CAT-Gen: Improving Robustness In NLP Models Via Controlled Adversarial Text Generation Highlight: In this work, we present a Controlled Adversarial Text Generation (CAT-Gen) model that, given an input text, generates adversarial texts through controllable attributes that are known to be invariant to task labels. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Tianlu Wang; Xuezhi Wang; Yao Qin; Ben Packer; Kang Li; Jilin Chen; Alex Beutel; Ed Chi; | |
418 | Seq2Edits: Sequence Transduction Using Span-level Edit Operations Highlight: We propose Seq2Edits, an open-vocabulary approach to sequence editing for natural language processing (NLP) tasks with a high degree of overlap between input and output texts. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Felix Stahlberg; Shankar Kumar; | |
419 | Controllable Meaning Representation To Text Generation: Linearization And Data Augmentation Strategies Highlight: We study the degree to which neural sequence-to-sequence models exhibit fine-grained controllability when performing natural language generation from a meaning representation. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Chris Kedzie; Kathleen McKeown; | |
420 | Blank Language Models Highlight: We propose Blank Language Model (BLM), a model that generates sequences by dynamically creating and filling in blanks. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Tianxiao Shen; Victor Quach; Regina Barzilay; Tommi Jaakkola; | |
421 | COD3S: Diverse Generation With Discrete Semantic Signatures Highlight: We present COD3S, a novel method for generating semantically diverse sentences using neural sequence-to-sequence (seq2seq) models. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Nathaniel Weir; João Sedoc; Benjamin Van Durme; | |
422 | Automatic Extraction Of Rules Governing Morphological Agreement Highlight: In this paper, we take steps towards automating this process by devising an automated framework for extracting a first-pass grammatical specification from raw text in a concise, human- and machine-readable format. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Aditi Chaudhary; Antonios Anastasopoulos; Adithya Pratapa; David R. Mortensen; Zaid Sheikh; Yulia Tsvetkov; Graham Neubig; | code |
423 | Tackling The Low-resource Challenge For Canonical Segmentation Highlight: We explore two new models for the task, borrowing from the closely related area of morphological generation: an LSTM pointer-generator and a sequence-to-sequence model with hard monotonic attention trained with imitation learning. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Manuel Mager; Özlem Çetinoğlu; Katharina Kann; | |
424 | IGT2P: From Interlinear Glossed Texts To Paradigms Highlight: We introduce a new task that speeds this process and automatically generates new morphological resources for natural language processing systems: IGT-to-paradigms (IGT2P). Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Sarah Moeller; Ling Liu; Changbing Yang; Katharina Kann; Mans Hulden; | |
425 | A Computational Approach To Understanding Empathy Expressed In Text-Based Mental Health Support Highlight: In this work, we present a computational approach to understanding how empathy is expressed in online mental health platforms. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Ashish Sharma; Adam Miner; David Atkins; Tim Althoff; | |
426 | Modeling Protagonist Emotions For Emotion-Aware Storytelling Highlight: In this paper, we present the first study on modeling the emotional trajectory of the protagonist in neural storytelling. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Faeze Brahman; Snigdha Chaturvedi; | |
427 | Help! Need Advice On Identifying Advice Highlight: We present preliminary models showing that while pre-trained language models are able to capture advice better than rule-based systems, advice identification is challenging, and we identify directions for future research. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Venkata Subrahmanyan Govindarajan; Benjamin Chen; Rebecca Warholic; Katrin Erk; Junyi Jessy Li; | |
428 | Quantifying Intimacy In Language Highlight: Here, we introduce a new computational framework for studying expressions of the intimacy in language with an accompanying dataset and deep learning model for accurately predicting the intimacy level of questions (Pearson r = 0.87). Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Jiaxin Pei; David Jurgens; | |
429 | Writing Strategies For Science Communication: Data And Computational Analysis Highlight: We compile a set of writing strategies drawn from a wide range of prescriptive sources and develop an annotation scheme allowing humans to recognize them. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Tal August; Lauren Kim; Katharina Reinecke; Noah A. Smith; | |
430 | Weakly Supervised Subevent Knowledge Acquisition Highlight: Acknowledging the scarcity of subevent knowledge, we propose a weakly supervised approach to extract subevent relation tuples from text and build the first large scale subevent knowledge base. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Wenlin Yao; Zeyu Dai; Maitreyi Ramaswamy; Bonan Min; Ruihong Huang; | |
431 | Biomedical Event Extraction As Sequence Labeling Highlight: We introduce Biomedical Event Extraction as Sequence Labeling (BeeSL), a joint end-to-end neural information extraction model. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Alan Ramponi; Rob van der Goot; Rosario Lombardo; Barbara Plank; | |
432 | Annotating Temporal Dependency Graphs Via Crowdsourcing Highlight: We present the construction of a corpus of 500 Wikinews articles annotated with temporal dependency graphs (TDGs) that can be used to train systems to understand temporal relations in text. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Jiarui Yao; Haoling Qiu; Bonan Min; Nianwen Xue; | |
433 | Introducing A New Dataset For Event Detection In Cybersecurity Texts Highlight: In particular, to facilitate the future research, we introduce a new dataset for this problem, characterizing the manual annotation for 30 important cybersecurity event types and a large dataset size to develop deep learning models. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Hieu Man Duc Trong; Duc Trong Le; Amir Pouran Ben Veyseh; Thuat Nguyen; Thien Huu Nguyen; | |
434 | CHARM: Inferring Personal Attributes From Conversations Highlight: This paper overcomes this limitation by devising CHARM: a zero-shot learning method that creatively leverages keyword extraction and document retrieval in order to predict attribute values that were never seen during training. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Anna Tigunova; Andrew Yates; Paramita Mirza; Gerhard Weikum; | |
435 | Event Detection: Gate Diversity And Syntactic Importance Scores For Graph Convolution Neural Networks Highlight: In this study, we propose a novel gating mechanism to filter noisy information in the hidden vectors of the GCN models for ED based on the information from the trigger candidate. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Viet Dac Lai; Tuan Ngo Nguyen; Thien Huu Nguyen; | |
436 | Severing The Edge Between Before And After: Neural Architectures For Temporal Ordering Of Events Highlight: In this paper, we propose a neural architecture and a set of training methods for ordering events by predicting temporal relations. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Miguel Ballesteros; Rishita Anubhai; Shuai Wang; Nima Pourdamghani; Yogarshi Vyas; Jie Ma; Parminder Bhatia; Kathleen McKeown; Yaser Al-Onaizan; | |
437 | How Much Knowledge Can You Pack Into The Parameters Of A Language Model? Highlight: In this short paper, we measure the practical utility of this approach by fine-tuning pre-trained models to answer questions without access to any external context or knowledge. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Adam Roberts; Colin Raffel; Noam Shazeer; | |
438 | EXAMS: A Multi-subject High School Examinations Dataset For Cross-lingual And Multilingual Question Answering Highlight: We propose EXAMS – a new benchmark dataset for cross-lingual and multilingual question answering for high school examinations. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Momchil Hardalov; Todor Mihaylov; Dimitrina Zlatkova; Yoan Dinkov; Ivan Koychev; Preslav Nakov; | code |
439 | End-to-End Synthetic Data Generation For Domain Adaptation Of Question Answering Systems Highlight: We propose an end-to-end approach for synthetic QA data generation. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Siamak Shakeri; Cicero Nogueira dos Santos; Henghui Zhu; Patrick Ng; Feng Nan; Zhiguo Wang; Ramesh Nallapati; Bing Xiang; | |
440 | Multi-Stage Pre-training For Low-Resource Domain Adaptation Highlight: We show that extending the vocabulary of the LM with domain-specific terms leads to further gains. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Rong Zhang; Revanth Gangi Reddy; Md Arafat Sultan; Vittorio Castelli; Anthony Ferritto; Radu Florian; Efsun Sarioglu Kayi; Salim Roukos; Avi Sil; Todd Ward; | |
441 | ISAAQ – Mastering Textbook Questions With Pre-trained Transformers And Bottom-Up And Top-Down Attention Highlight: For the first time, this paper taps on the potential of transformer language models and bottom-up and top-down attention to tackle the language and visual understanding challenges this task entails. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Jose Manuel Gomez-Perez; Raúl Ortega; | |
442 | SubjQA: A Dataset For Subjectivity And Review Comprehension Highlight: We find that subjectivity is an important feature in the case of QA, albeit with more intricate interactions between subjectivity and QA performance than found in previous work on sentiment analysis. We develop a new dataset which allows us to investigate this relationship. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Johannes Bjerva; Nikita Bhutani; Behzad Golshan; Wang-Chiew Tan; Isabelle Augenstein; | code |
443 | Widget Captioning: Generating Natural Language Description For Mobile User Interface Elements Highlight: We propose widget captioning, a novel task for automatically generating language descriptions for UI elements from multimodal input including both the image and the structural representations of user interfaces. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Yang Li; Gang Li; Luheng He; Jingjie Zheng; Hong Li; Zhiwei Guan; | |
444 | Unsupervised Natural Language Inference Via Decoupled Multimodal Contrastive Learning Highlight: In this paper, we propose Multimodal Aligned Contrastive Decoupled learning (MACD) network. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Wanyun Cui; Guangyu Zheng; Wei Wang; | |
445 | Digital Voicing Of Silent Speech Highlight: In this paper, we consider the task of digitally voicing silent speech, where silently mouthed words are converted to audible speech based on electromyography (EMG) sensor measurements that capture muscle impulses. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
David Gaddy; Dan Klein; | |
446 | Imitation Attacks And Defenses For Black-box Machine Translation Systems Highlight: To mitigate these vulnerabilities, we propose a defense that modifies translation outputs in order to misdirect the optimization of imitation models. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Eric Wallace; Mitchell Stern; Dawn Song; | |
447 | Sequence-Level Mixed Sample Data Augmentation Highlight: This work proposes a simple data augmentation approach to encourage compositional behavior in neural models for sequence-to-sequence problems. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Demi Guo; Yoon Kim; Alexander Rush; | |
448 | Consistency Of A Recurrent Language Model With Respect To Incomplete Decoding Highlight: Based on these insights, we propose two remedies which address inconsistency: consistent variants of top-k and nucleus sampling, and a self-terminating recurrent language model. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Sean Welleck; Ilia Kulikov; Jaedeok Kim; Richard Yuanzhe Pang; Kyunghyun Cho; | |
449 | An Exploration Of Arbitrary-Order Sequence Labeling Via Energy-Based Inference Networks Highlight: In this work, we propose several high-order energy terms to capture complex dependencies among labels in sequence labeling, including several that consider the entire label sequence. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Lifu Tu; Tianyu Liu; Kevin Gimpel; | |
450 | Ensemble Distillation For Structured Prediction: Calibrated, Accurate, Fast—Choose Three Highlight: In this paper, we study \textit{ensemble distillation} as a general framework for producing well-calibrated structured prediction models while avoiding the prohibitive inference-time cost of ensembles. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Steven Reich; David Mueller; Nicholas Andrews; | |
451 | Inducing Target-Specific Latent Structures For Aspect Sentiment Classification Highlight: We propose gating mechanisms to dynamically combine information from word dependency graphs and latent graphs which are learned by self-attention networks. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Chenhua Chen; Zhiyang Teng; Yue Zhang; | |
452 | Affective Event Classification With Discourse-enhanced Self-training Highlight: Our research introduces new classification models to assign affective polarity to event phrases. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Yuan Zhuang; Tianyu Jiang; Ellen Riloff; | |
453 | Deep Weighted MaxSAT For Aspect-based Opinion Extraction Highlight: We adopt the MaxSAT semantics to model logic inference process and smoothly incorporate a weighted version of MaxSAT that connects deep neural networks and a graphical model in a joint framework. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Meixi Wu; Wenya Wang; Sinno Jialin Pan; | |
454 | Multi-view Story Characterization From Movie Plot Synopses And Reviews Highlight: This paper considers the problem of characterizing stories by inferring properties such as theme and style using written synopses and reviews of movies. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Sudipta Kar; Gustavo Aguilar; Mirella Lapata; Thamar Solorio; | code |
455 | Mind Your Inflections! Improving NLP For Non-Standard Englishes With Base-Inflection Encoding Highlight: We propose Base-Inflection Encoding (BITE), a method to tokenize English text by reducing inflected words to their base forms before reinjecting the grammatical information as special symbols. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Samson Tan; Shafiq Joty; Lav Varshney; Min-Yen Kan; | |
456 | Measuring The Similarity Of Grammatical Gender Systems By Comparing Partitions Highlight: To quantify the similarity, we define gender systems extensionally, thereby reducing the problem of comparisons between languages’ gender systems to cluster evaluation. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Arya D. McCarthy; Adina Williams; Shijia Liu; David Yarowsky; Ryan Cotterell; | |
457 | RethinkCWS: Is Chinese Word Segmentation A Solved Task? Highlight: In this paper, we take stock of what we have achieved and rethink what’s left in the CWS task. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Jinlan Fu; Pengfei Liu; Qi Zhang; Xuanjing Huang; | code |
458 | Learning To Pronounce Chinese Without A Pronunciation Dictionary Highlight: We demonstrate a program that learns to pronounce Chinese text in Mandarin, without a pronunciation dictionary. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Christopher Chu; Scot Fang; Kevin Knight; | |
459 | Dynamic Anticipation And Completion For Multi-Hop Reasoning Over Sparse Knowledge Graph Highlight: To solve these problems, we propose a multi-hop reasoning model over sparse KGs, by applying novel dynamic anticipation and completion strategies: (1) The anticipation strategy utilizes the latent prediction of embedding-based models to make our model perform more potential path search over sparse KGs. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Xin Lv; Xu Han; Lei Hou; Juanzi Li; Zhiyuan Liu; Wei Zhang; Yichi Zhang; Hao Kong; Suhui Wu; | code |
460 | Knowledge Association With Hyperbolic Knowledge Graph Embeddings Highlight: We propose a hyperbolic relational graph neural network for KG embedding and capture knowledge associations with a hyperbolic transformation. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Zequn Sun; Muhao Chen; Wei Hu; Chengming Wang; Jian Dai; Wei Zhang; | |
461 | Domain Knowledge Empowered Structured Neural Net For End-to-End Event Temporal Relation Extraction Highlight: To address these issues, we propose a framework that enhances deep neural network with distributional constraints constructed by probabilistic domain knowledge. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Rujun Han; Yichao Zhou; Nanyun Peng; | |
462 | TeMP: Temporal Message Passing For Temporal Knowledge Graph Completion Highlight: We propose the Temporal Message Passing (TeMP) framework to address these challenges by combining graph neural networks, temporal dynamics models, data imputation and frequency-based gating techniques. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Jiapeng Wu; Meng Cao; Jackie Chi Kit Cheung; William L. Hamilton; | |
463 | Understanding The Difficulty Of Training Transformers Highlight: Our objective here is to understand {\_}{\_}what complicates Transformer training{\_}{\_} from both empirical and theoretical perspectives. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Liyuan Liu; Xiaodong Liu; Jianfeng Gao; Weizhu Chen; Jiawei Han; | |
464 | An Empirical Study Of Generation Order For Machine Translation Highlight: In this work, we present an empirical study of generation order for machine translation. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
William Chan; Mitchell Stern; Jamie Kiros; Jakob Uszkoreit; | |
465 | Inference Strategies For Machine Translation With Conditional Masking Highlight: We identify a thresholding strategy that has advantages over the standard mask-predict algorithm, and provide analyses of its behavior on machine translation tasks. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Julia Kreutzer; George Foster; Colin Cherry; | |
466 | AmbigQA: Answering Ambiguous Open-domain Questions Highlight: In this paper, we introduce AmbigQA, a new open-domain question answering task which involves finding every plausible answer, and then rewriting the question for each one to resolve the ambiguity. Related Papers Related Patents Related Grants Related Orgs Related Experts Details |
Sewon Min; Julian Michael; Hannaneh Hajishirzi; Luke Zettlemoyer; | code |
467 | Tell Me How To Ask Again: Question Data Augmentation With Controllable Rewriting In Continuous Space Highlight: In this paper, we propose a novel data augmentation method, referred to as Controllable Rewriting based Question Data Augmentation (CRQDA), for machine reading comprehension (MRC), question generation, a |