Joint Semantic Analysis with Document-Level Cross-Task Coherence Rewards

Coreference resolution and semantic role labeling are NLP tasks that capture different aspects of semantics, indicating respectively, which expressions refer to the same entity, and what semantic roles expressions serve in the sentence. However, they are often closely interdependent, and both generally necessitate natural language understanding. Do they form a coherent abstract representation of documents? We present a neural network architecture for joint coreference resolution and semantic role labeling for English, and train graph neural networks to model the 'coherence' of the combined shallow semantic graph. Using the resulting coherence score as a reward for our joint semantic analyzer, we use reinforcement learning to encourage global coherence over the document and between semantic annotations. This leads to improvements on both tasks in multiple datasets from different domains, and across a range of encoders of different expressivity, calling, we believe, for a more holistic approach to semantics in NLP.

[1]  Luke S. Zettlemoyer,et al.  Deep Semantic Role Labeling: What Works and What’s Next , 2017, ACL.

[2]  Max Welling,et al.  Semi-Supervised Classification with Graph Convolutional Networks , 2016, ICLR.

[3]  Chen Qiu,et al.  Rewarding Coreference Resolvers for Being Consistent with World Knowledge , 2019, EMNLP/IJCNLP.

[4]  Andrew McCallum,et al.  Linguistically-Informed Self-Attention for Semantic Role Labeling , 2018, EMNLP.

[5]  Daniel Jurafsky,et al.  Semantic Role Chunking Combining Complementary Syntactic Views , 2005, CoNLL.

[6]  Uwe Reyle,et al.  From Discourse to Logic - Introduction to Modeltheoretic Semantics of Natural Language, Formal Logic and Discourse Representation Theory , 1993, Studies in linguistics and philosophy.

[7]  Simone Paolo Ponzetto,et al.  Exploiting Semantic Role Labeling, WordNet and Wikipedia for Coreference Resolution , 2006, NAACL.

[8]  Ari Rappoport,et al.  A Transition-Based Directed Acyclic Graph Parser for UCCA , 2017, ACL.

[9]  Dan Roth,et al.  The Importance of Syntactic Parsing and Inference in Semantic Role Labeling , 2008, CL.

[10]  Guokun Lai,et al.  RACE: Large-scale ReAding Comprehension Dataset From Examinations , 2017, EMNLP.

[11]  Ting Liu,et al.  Learning Semantic Representations of Users and Products for Document Level Sentiment Classification , 2015, ACL.

[12]  Rico Sennrich,et al.  Context-Aware Neural Machine Translation Learns Anaphora Resolution , 2018, ACL.

[13]  Udo Kruschwitz,et al.  Phrase detectives: Utilizing collective intelligence for internet-scale language resource creation , 2013, TIIS.

[14]  Zhen-Hua Ling,et al.  Distraction-Based Neural Networks for Document Summarization , 2016, ArXiv.

[15]  Ming-Wei Chang,et al.  Well-Read Students Learn Better: On the Importance of Pre-training Compact Models , 2019 .

[16]  Vincent Ng,et al.  Shallow Semantics for Coreference Resolution , 2007, IJCAI.

[17]  Jimmy Ba,et al.  Adam: A Method for Stochastic Optimization , 2014, ICLR.

[18]  Yuchen Zhang,et al.  CoNLL-2012 Shared Task: Modeling Multilingual Unrestricted Coreference in OntoNotes , 2012, EMNLP-CoNLL Shared Task.

[19]  Marcin Junczys-Dowmunt,et al.  Microsoft Translator at WMT 2019: Towards Large-Scale Document-Level Neural Machine Translation , 2019, WMT.

[20]  Parminder Bhatia,et al.  Better Document-level Sentiment Analysis from RST Discourse Parsing , 2015, EMNLP.

[21]  Ari Rappoport,et al.  The State of the Art in Semantic Representation , 2017, ACL.

[22]  Dan Klein,et al.  Simple Coreference Resolution with Rich Syntactic and Semantic Features , 2009, EMNLP.

[23]  Lou Boves,et al.  Discourse-based answering of why-questions , 2006, Trait. Autom. des Langues.

[24]  Simone Paolo Ponzetto,et al.  Semantic Role Labeling for Coreference Resolution , 2006, EACL.

[25]  Luke S. Zettlemoyer,et al.  End-to-end Neural Coreference Resolution , 2017, EMNLP.

[26]  Johan Bos,et al.  The Groningen Meaning Bank , 2013, JSSP.

[27]  Ming-Wei Chang,et al.  BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding , 2019, NAACL.

[28]  Omri Abend,et al.  On the Weaknesses of Reinforcement Learning for Neural Machine Translation , 2019, ICLR.

[29]  Alexander M. Rush,et al.  Antecedent Prediction Without a Pipeline , 2016, CORBON@HLT-NAACL.

[30]  Yang Liu,et al.  Learning Structured Text Representations , 2017, TACL.

[31]  Jackie Chi Kit Cheung,et al.  A Generalized Knowledge Hunting Framework for the Winograd Schema Challenge , 2018, NAACL-HLT.

[32]  Yan Song,et al.  Incorporating Context and External Knowledge for Pronoun Coreference Resolution , 2019, NAACL.

[33]  John Langford,et al.  Relating reinforcement learning performance to classification performance , 2005, ICML '05.

[34]  Beth Levin,et al.  English Verb Classes and Alternations: A Preliminary Investigation , 1993 .

[35]  Martha Palmer,et al.  AMR Beyond the Sentence: the Multi-sentence AMR corpus , 2018, COLING.

[36]  Kadri Hacioglu,et al.  Semantic Role Labeling Using Dependency Trees , 2004, COLING.

[37]  Omer Levy,et al.  BERT for Coreference Resolution: Baselines and Analysis , 2019, EMNLP/IJCNLP.

[38]  Ari Rappoport,et al.  Universal Conceptual Cognitive Annotation (UCCA) , 2013, ACL.

[39]  Luke S. Zettlemoyer,et al.  Higher-Order Coreference Resolution with Coarse-to-Fine Inference , 2018, NAACL.

[40]  Christopher D. Manning,et al.  Deep Reinforcement Learning for Mention-Ranking Coreference Models , 2016, EMNLP.

[41]  Xavier Carreras,et al.  Introduction to the CoNLL-2005 Shared Task: Semantic Role Labeling , 2005, CoNLL.

[42]  Mirella Lapata,et al.  Discourse Representation Structure Parsing , 2018, ACL.

[43]  Jeffrey Pennington,et al.  GloVe: Global Vectors for Word Representation , 2014, EMNLP.

[44]  Sadao Kurohashi,et al.  Entity-Centric Joint Modeling of Japanese Coreference Resolution and Predicate Argument Structure Analysis , 2018, ACL.

[45]  Christopher D. Manning,et al.  Improving Coreference Resolution by Learning Entity-Level Distributed Representations , 2016, ACL.

[46]  Omer Levy,et al.  Jointly Predicting Predicates and Arguments in Neural Semantic Role Labeling , 2018, ACL.

[47]  Samuel R. Bowman,et al.  A Gold Standard Dependency Corpus for English , 2014, LREC.

[48]  Vincent Ng,et al.  Coreference Resolution with World Knowledge , 2011, ACL.

[49]  Pietro Liò,et al.  Deep Graph Infomax , 2018, ICLR.

[50]  Luo Si,et al.  Syntax-Aware Neural Semantic Role Labeling , 2019, AAAI.

[51]  Philipp Koehn,et al.  Abstract Meaning Representation for Sembanking , 2013, LAW@ACL.

[52]  Xinya Du,et al.  Be Consistent! Improving Procedural Text Comprehension using Label Consistency , 2019, NAACL.

[53]  Andrew McCallum,et al.  Joint Parsing and Semantic Role Labeling , 2005, CoNLL.

[54]  Nathan Schneider,et al.  Semantically Constrained Multilayer Annotation: The Case of Coreference , 2019, Proceedings of the First International Workshop on Designing Meaning Representations.

[55]  Wai Lam,et al.  AMR Parsing via Graph-Sequence Iterative Inference , 2020, ACL.

[56]  Philippe Langlais,et al.  WikiCoref: An English Coreference-annotated Corpus of Wikipedia Articles , 2016, LREC.

[57]  Jieyu Zhao,et al.  Gender Bias in Coreference Resolution: Evaluation and Debiasing Methods , 2018, NAACL.

[58]  Gholamreza Haffari,et al.  Document Context Neural Machine Translation with Memory Networks , 2017, ACL.

[59]  Hong Chen,et al.  PreCo: A Large-scale Dataset in Preschool Vocabulary for Coreference Resolution , 2018, EMNLP.

[60]  Farid Meziane,et al.  A Discourse-Based Approach for Arabic Question Answering , 2016, ACM Trans. Asian Low Resour. Lang. Inf. Process..