Conversational Response Re-ranking Based on Event Causality and Role Factored Tensor Event Embedding

Katsuhito Sudoh; Koichiro Yoshino; Satoshi Nakamura; Shohei Tanaka

arxiv: 1906.09795 · v1 · pith:N4BEFDT2new · submitted 2019-06-24 · 💻 cs.CL

Conversational Response Re-ranking Based on Event Causality and Role Factored Tensor Event Embedding

Shohei Tanaka , Koichiro Yoshino , Katsuhito Sudoh , Satoshi Nakamura This is my paper

Pith reviewed 2026-05-25 17:43 UTC · model grok-4.3

classification 💻 cs.CL

keywords response re-rankingevent causalityrole factored tensor modeldialogue systemsconversational response selectionevent embeddingdialogue coherencydialogue continuity

0 comments

The pith

Re-ranking conversational responses by matching event causality relations yields more coherent and continuous dialogues.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper introduces a re-ranking method that selects responses by checking causal connections between events mentioned in the dialogue history and those in candidate replies. It represents events through the Role Factored Tensor Model so that matching can succeed even when the system knows only a limited set of causal relations. Experiments on generated candidates show gains in human-judged coherency and in how well the replies keep the conversation flowing. A reader would care because many dialogue systems produce replies that feel logically disconnected from what came before. The approach leaves the original response generator unchanged and only reorders its outputs.

Core claim

The paper claims that re-ranking response candidates generated from conversational models, using event causality relations encoded in Role Factored Tensor Model embeddings, produces system responses with measurably higher coherency and dialogue continuity.

What carries the argument

Role Factored Tensor Model event embeddings, which factor participant roles to produce distributed representations that support robust matching of causal event pairs.

If this is right

Responses maintain greater logical consistency with events already stated in the history.
Dialogue continuity rises because replies more reliably follow plausible causal chains.
The re-ranking works without requiring exhaustive prior knowledge of all possible event relations.
The method applies to any pool of candidate replies produced by an existing conversational model.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

The same causality filter could be tested on longer multi-turn conversations to check whether continuity gains accumulate across turns.
Replacing the tensor model with a simpler embedding method would isolate how much the role factoring itself contributes to the matching robustness.
The technique might transfer to other text-generation settings where maintaining causal consistency between sentences matters.

Load-bearing premise

The Role Factored Tensor Model supplies robust matching of event causality relations even when the system has only limited event causality knowledge.

What would settle it

A new experiment on held-out dialogues in which human raters assign the re-ranked responses no higher coherency or continuity scores than the unranked candidates would falsify the reported improvement.

Figures

Figures reproduced from arXiv: 1906.09795 by Katsuhito Sudoh, Koichiro Yoshino, Satoshi Nakamura, Shohei Tanaka.

**Figure 1.** Figure 1: Neural conversational model+re-ranking using event causality; a response that has an event causality relation (“be exhausted” → “relax”) to the dialogue history is selected by the re-ranking. predicate 1 argument 1 predicate 2 argument 2 lif t be stressed out - relieve stress 10.02 [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗

**Figure 2.** Figure 2: Model architecture of predicate embedding [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗

read the original abstract

We propose a novel method for selecting coherent and diverse responses for a given dialogue context. The proposed method re-ranks response candidates generated from conversational models by using event causality relations between events in a dialogue history and response candidates (e.g., ``be stressed out'' precedes ``relieve stress''). We use distributed event representation based on the Role Factored Tensor Model for a robust matching of event causality relations due to limited event causality knowledge of the system. Experimental results showed that the proposed method improved coherency and dialogue continuity of system responses.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Desk Editor's Note private letter to a colleague

The paper adds event-causality re-ranking via role-factored tensor embeddings to dialogue response selection, but supplies no direct test of the claimed robustness under sparse knowledge.

read the letter

The main takeaway is a re-ranking step that pulls event causality links from dialogue history to score response candidates, with the Role Factored Tensor Model meant to keep the matching reliable even when the system has incomplete causality data. The abstract frames this as the novel piece, and the motivation for using tensor factorization on event roles is laid out clearly enough to follow without extra background. The approach itself is a modest but concrete engineering move that could slot into existing generation pipelines to favor responses that continue prior events more naturally. That part is straightforward and addresses a real pain point in keeping dialogues coherent over turns. The experiments are said to show gains in coherency and continuity, which is the kind of outcome that matters for deployed systems. The soft spot is exactly the one flagged in the stress test. Nothing in the provided description isolates whether the tensor model actually compensates for missing causality relations or simply rides on whatever coverage happens to be in the training data. There are also no reported baselines, metrics, data splits, or statistical checks, so the size of the improvement and its reliability stay hard to judge. If the full paper contains ablations on sparsity or comparisons against simpler matching, that would close the gap; otherwise the central robustness claim rests on assertion rather than evidence. This is the sort of incremental method paper that dialogue-system groups might want to look at for implementation ideas. It is narrow enough that it will not reshape the field, but the idea is reproducible enough that a serious referee could usefully check the experimental controls and ask for the missing ablations. I would send it to review rather than desk-reject.

Referee Report

2 major / 2 minor

Summary. The paper proposes a response re-ranking method for dialogue systems that leverages event causality relations between dialogue history and candidate responses, using Role Factored Tensor Model embeddings to enable robust matching despite limited system knowledge of causality. It claims that this approach improves coherency and dialogue continuity, as shown by experimental results.

Significance. If the experimental claims are substantiated, the work could contribute to more coherent open-domain dialogue by incorporating causal structure in a manner tolerant to sparse knowledge; the tensor factorization approach for events is a potentially useful technical element if its robustness benefit is isolated.

major comments (2)

[Experimental Results] Experimental Results section: no information is supplied on the baselines, evaluation metrics for coherency/continuity, statistical significance tests, or train/test splits, preventing assessment of the claimed improvements.
[Abstract and Method] Abstract and §3 (Method): the central claim that the Role Factored Tensor Model supplies robust causality matching under limited knowledge is not supported by any ablation, sparsity analysis, or non-robust baseline comparison that would isolate this property.

minor comments (2)

[Abstract] The abstract would benefit from a brief statement of the dataset size and primary metric to allow readers to gauge the scale of the experiments.
[Method] Notation for the tensor factorization (e.g., role vectors and event composition) could include a short worked example for clarity.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the constructive feedback. We address the two major comments below and will revise the manuscript to provide the requested details and analyses.

read point-by-point responses

Referee: [Experimental Results] Experimental Results section: no information is supplied on the baselines, evaluation metrics for coherency/continuity, statistical significance tests, or train/test splits, preventing assessment of the claimed improvements.

Authors: We agree that the Experimental Results section lacks these details. In the revised manuscript we will add a complete description of the baselines, the precise metrics used to measure coherency and dialogue continuity, statistical significance test results, and the train/test split procedure. revision: yes
Referee: [Abstract and Method] Abstract and §3 (Method): the central claim that the Role Factored Tensor Model supplies robust causality matching under limited knowledge is not supported by any ablation, sparsity analysis, or non-robust baseline comparison that would isolate this property.

Authors: The manuscript asserts robustness from the tensor factorization but does not isolate this benefit via ablation or sparsity experiments. We will add an ablation study and a sparsity analysis comparing the Role Factored Tensor Model against non-robust baselines to support the claim. revision: yes

Circularity Check

0 steps flagged

No circularity: derivation relies on external tensor model and data without self-referential reduction

full rationale

The paper describes a re-ranking approach that applies an existing Role Factored Tensor Model to match event causality relations between dialogue history and candidate responses. The abstract and provided text contain no equations, parameter-fitting steps, or self-citations that would make any claimed prediction equivalent to its inputs by construction. The central improvement in coherency is presented as an experimental outcome rather than a mathematical identity or renamed fit. No load-bearing premise reduces to a self-definition or imported uniqueness theorem from the same authors. The derivation chain therefore remains independent of the target result.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

Abstract provides no identifiable free parameters, axioms, or invented entities; full text would be required to audit these.

pith-pipeline@v0.9.0 · 5621 in / 842 out tokens · 38467 ms · 2026-05-25T17:43:54.970416+00:00 · methodology

discussion (0)

Lean theorems connected to this paper

Citations machine-checked in the Pith Canon. Every link opens the source theorem in the public Lean library.

IndisputableMonolith/Cost/FunctionalEquation.lean washburn_uniqueness_aczel unclear

?

unclear
Relation between the paper passage and the cited Recognition theorem.

We use distributed event representation based on the Role Factored Tensor Model (RFTM) ... e = ∑_a W_a T(v_p, v_a)
IndisputableMonolith/Foundation/RealityFromDistinction.lean reality_from_one_distinction unclear

?

unclear
Relation between the paper passage and the cited Recognition theorem.

liftemb(eh, er) = lift(ec, ee) * mean(sim(eh, ec), sim(er, ee))

What do these tags mean?

matches: The paper's claim is directly supported by a theorem in the formal canon.
supports: The theorem supports part of the paper's argument, but the paper may add assumptions or extra steps.
extends: The paper goes beyond the formal theorem; the theorem is a base layer rather than the whole result.
uses: The paper appears to rely on the theorem as machinery.
contradicts: The paper's claim conflicts with a theorem or certificate in the canon.
unclear: Pith found a possible connection, but the passage is too broad, indirect, or ambiguous to say the theorem truly supports the claim.

Reference graph

Works this paper leans on

35 extracted references · 35 canonical work pages · 2 internal anchors

[1]

URL: " 'urlintro :=

ENTRY address author booktitle chapter edition editor howpublished institution journal key month note number organization pages publisher school series title type volume year eprint doi pubmed url lastchecked label extra.label sort.label short.list INTEGERS output.state before.all mid.sentence after.sentence after.block STRINGS urlintro eprinturl eprintpr...

work page
[2]

write newline

" write newline "" before.all 'output.state := FUNCTION n.dashify 't := "" t empty not t #1 #1 substring "-" = t #1 #2 substring "--" = not "--" * t #2 global.max substring 't := t #1 #1 substring "-" = "-" * t #2 global.max substring 't := while if t #1 #1 substring * t #2 global.max substring 't := if while FUNCTION word.in bbl.in capitalize " " * FUNCT...

work page
[3]

Dzmitry Bahdanau, Kyunghyunand Cho, and Yoshua Bengio. 2015. Neural Machine Translation by Jointly Learning to Align and Translate . In Proceedings of the 3rd International Conference on Learning Representations (ICLR)

work page 2015
[4]

Dasha Bogdanova and Jennifer Foster. 2016. This is how we do it: Answer Reranking for Open-Domain How Questions with Paragraph Vectors and Minimal Feature Engineering . In Proceedings of the 15th Annual Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (NAACL-HLT), pages 1290--1295

work page 2016
[5]

Kyunghyun Cho, Bart van Merrienboer, Caglar Gulcehre, Dzmitry Bahdanau, Fethi Bougares, Holger Schwenk, and Yoshua Bengio. 2014. Learning Phrase Representations Using RNN Encoder-Decoder for Statistical Machine Translation . In Proceedings of the 2014 Conference on Empirical Methods in Natural Language Processing (EMNLP)

work page 2014
[6]

Junyoung Chung, Caglar Gulcehre, KyungHyun Cho, and Yoshua Bengio. 2014. Empirical Evaluation of Gated Recurrent Neural Networks on Sequence Modeling . In Proceedings of the 28th Conference Neural Information Processing Systems, Deep Learning and Representation Learning Workshop (NIPS)

work page 2014
[7]

George Doddington. 2002. Automatic Evaluation of Machine Translation Quality Using N-gram Co-occurrence Statistics . In Proceedings of the 2nd International Conference on Human Language Technology Research (HLT), pages 138--145

work page 2002
[8]

Motoyasu Fujita, Rafal Rzepka, and Kenji Araki. 2011. Evaluation of Utterances Based on Causal Knowledge Retrieved from Blogs . In Proceedings of the 14th IASTED International Conference Artificial Intelligence and Soft Computing (ASC), pages 294--299

work page 2011
[9]

Forgues Gabriel, Joelle Pineau, Jean-Marie Larchev \^ e que, and R \' e al Tremblay. 2014. Bootstrapping Dialog Systems with Word Embeddings

work page 2014
[10]

Peter Jansen, Mihai Surdeanu, and Peter Clark. 2014. Discourse Complements Lexical Semantics for Non-factoid Answer Reranking . In Proceedings of the 52nd Annual Meeting of the Association for Computational Linguistics (ACL), pages 977--986

work page 2014
[11]

Daisuke Kawahara and Sadao Kurohashi. 2006. A Fully-Lexicalized Probabilistic Model for Japanese Syntactic and Case Structure Analysis . In Proceedings of Human Language Technology Conference - North American Chapter of the Association for Computational Linguistics Annual Meeting (HLT-NAACL), pages 176--183

work page 2006
[12]

Kingma and Jimmy Ba

Diederik P. Kingma and Jimmy Ba. 2015. Adam: A Method for Stochastic Optimization . In Proceedings of the 3rd International Conference on Learning Representations (ICLR)

work page 2015
[13]

Taku Kudo and John Richardson. 2018. SentencePiece: A Simple and Language Independent Subword Tokenizer and Detokenizer for Neural Text Processing . In Proceedings of the 2018 Conference on Empirical Methods in Natural Language Processing (EMNLP)

work page 2018
[14]

Jiwei Li, Michel Galley, Chris Brockett, Jianfeng Gao, and Bill Dolan. 2016. A Diversity-Promoting Objective Function for Neural Conversation Models . In Proceedings of the 15th Annual Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (NAACL-HLT), pages 110--119

work page 2016
[15]

Serban, Michael Noseworthy, Laurent Charlin, and Joelle Pineau

Chia-Wei Liu, Ryan Lowe, Iulian V. Serban, Michael Noseworthy, Laurent Charlin, and Joelle Pineau. 2016. How NOT to Evaluate Your Dialogue System: An Empirical Study of Unsupervised Evaluation Metrics for Dialogue Response Generation . In Proceedings of the 2016 Conference on Empirical Methods in Natural Language Processing (EMNLP)

work page 2016
[16]

Minh-Thang Luong, Hieu Pham, and Christopher D. Manning. 2015. Effective Approaches to Attention-Based Neural Machine Translation . In Proceedings of the 2015 Conference on Empirical Methods in Natural Language Processing (EMNLP)

work page 2015
[17]

Wolfgang Macherey, Maxim Krikun, Yuan Cao, Qin Gao, Klaus Macherey, Jeff Klingner, Apurva Shah, Melvin Johnson, Xiaobing Liu, ukasz Kaiser, Stephan Gouws, Yoshikiyo Kato, Taku Kudo, Hideto Kazawa, Keith Stevens, George Kurian, Nishant Patil, Wei Wang, Cliff Young, Jason Smith, Jason Riesa, Alex Rudnick, Oriol Vinyals, Greg Corrado, Macduff Hughes, and Jef...

work page internal anchor Pith review Pith/arXiv arXiv 2016
[18]

Tomas Mikolov, Kai Chen, Greg Corrado, and Jeffrey Deany. 2013 a . Efficient Estimation of Word Representations in Vector Space . In Proceedings of the 1st International Conference on Learning Representations (ICLR)

work page 2013
[19]

Tomas Mikolov, Ilya Sutskever, Kai Chen, Greg Corrado, and Jeffrey Dean. 2013 b . Distributed Representations of Words and Phrases and Their Compositionality . In Proceedings of the 26th International Conference on Neural Information Processing Systems (NIPS), volume 2, pages 3111--3119

work page 2013
[20]

Tomas Mikolov, Wen-tau Yih, and Geoffrey Zweig. 2013 c . Linguistic Regularities in Continuous Space Word Representations . In Proceedings of the 12th Annual Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (NAACL-HLT), pages 746--751

work page 2013
[21]

Ryo Nakamura, Katsuhito Sudoh, Koichiro Yoshino, and Satoshi Nakamura. 2019. Another Diversity-Promoting Objective Function for Neural Dialogue Generation . In Proceedings of the 33rd Association for the Advancement of Artificial Intelligence Conference on Artificial Intelligence, Workshop on Reasoning and Learning for Human-Machine Dialogues (DEEP-DIAL 2...

work page 2019
[22]

David Newman, Jey Han Lau, Karl Grieser, and Timothy Baldwin. 2010. Automatic Evaluation of Topic Coherence . In Proceedings of the 11th Annual Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (NAACL-HLT), pages 100--108

work page 2010
[23]

Jong-Hoon Oh, Kentaro Torisawa, Chikara Hashimoto, Ryu Iida, Masahiro Tanaka, and Julien Kloetzer. 2016. A Semi-supervised Learning Approach to Why-Question Answering . In Proceedings of the 30th Association for the Advancement of Artificial Intelligence Conference on Artificial Intelligence (AAAI), pages 3022--3029

work page 2016
[24]

Jong-Hoon Oh, Kentaro Torisawa, Chikara Hashimoto, Motoki Sano, Stijn De Saeger, and Kiyonori Ohtake. 2013. Why-Question Answering Using Intra- and Inter-Sentential Causal Relations . In Proceedings of the 51st Annual Meeting of the Association for Computational Linguistics (ACL), pages 1733--1743

work page 2013
[25]

Jong-Hoon Oh, Kentaro Torisawa, Canasai Kruengkrai, Ryu Iida, and Julien Kloetzer. 2017. Multi-Column Convolutional Neural Networks with Causality-Attention for Why-Question Answering . In Proceedings of the 10th Association for Computing Machinery International Conference on Web Search and Data Mining (WSDM), pages 415--424

work page 2017
[26]

Junki Ohmura and Maxine Eskenazi. 2018. Context-Aware Dialog Re-ranking for Task-Oriented Dialog Systems . In Proceedings of IEEE Spoken Language Technology Workshop (SLT)

work page 2018
[27]

Kishore Papineni, Salim Roukos, Todd Ward, and Wei-Jing Zhu. 2002. BLEU: a Method for Automatic Evaluation of Machine Translation . In Proceedings of the 40th Annual Meeting of the Association for Computational Linguistics (ACL), pages 311--318

work page 2002
[28]

Ashwin Ram, Rohit Prasad, Chandra Khatri, Anu Venkatesh, Raefer Gabriel, Qing Liu, Jeff Nunn, Behnam Hedayatnia, Ming Cheng, Ashish Nagar, Eric King, Kate Bland, Amanda Wartick, Yi Pan, Han Song, Sk Jayadevan, Gene Hwang, and Art Pettigrue. 2018. Conversational AI: The Science Behind the Alexa Prize . In arXiv:1801.03604

work page internal anchor Pith review Pith/arXiv arXiv 2018
[29]

Ryohei Sasano and Sadao Kurohashi. 2011. A Discriminative Approach to Japanese Zero Anaphora Resolution with Large-Scale Lexicalized Case Frames . In Proceedings of the 5th International Joint Conference on Natural Language Processing (IJCNLP), pages 758--766

work page 2011
[30]

Serban, Alessandro Sordoni, Yoshua Bengio, Aaron Courville, and Joelle Pineau

Iulian V. Serban, Alessandro Sordoni, Yoshua Bengio, Aaron Courville, and Joelle Pineau. 2016. Building End-To-End Dialogue Systems Using Generative Hierarchical Neural Network Models . In Proceedings of the 30th Association for the Advancement of Artificial Intelligence Conference on Artificial Intelligence (AAAI)

work page 2016
[31]

Tomohide Shibata, Shotaro Kohama, and Sadao Kurohashi. 2014. A Large Scale Database of Strongly-Related Events in Japanese . In Proceedings of the 9th International Conference on Language Resources and Evalu ation (LREC)

work page 2014
[32]

Tomohide Shibata and Sadao Kurohashi. 2011. Acquiring Strongly-Related Events Using Predicate-Argument Co-occurring Statist ics and Case Frames . In Proceedings of the 5th International Joint Conference on Natural Language Proce ssing (IJCNLP), pages 1028--1036

work page 2011
[33]

Simonsen, and Jian-Yun Nie

Alessandro Sordoni, Yoshua Bengio, Hossein Vahabi, Christina Lioma, Jakob G. Simonsen, and Jian-Yun Nie. 2015. A Hierarchical Recurrent Encoder-Decoder For Generative Context-Aware Query Suggestion . In Proceedings of the 24th Association for Computing Machinery International Conference on Information Knowledge and Management (ACM)

work page 2015
[34]

Oriol Vinyals and Quoc V. Le. 2015. A Neural Conversational Model . In Proceedings of the 32nd International Conference on Machine Learning, Deep Learning Workshop (ICML)

work page 2015
[35]

Noah Weber, Niranjan Balasubramanian, and Nathanael Chambers. 2018. Event Representations with Tensor-Based Compositions . In Proceedings of the 32nd Association for the Advancement of Artificial Intelligence Conference on Artificial Intelligence (AAAI)

work page 2018

[1] [1]

URL: " 'urlintro :=

ENTRY address author booktitle chapter edition editor howpublished institution journal key month note number organization pages publisher school series title type volume year eprint doi pubmed url lastchecked label extra.label sort.label short.list INTEGERS output.state before.all mid.sentence after.sentence after.block STRINGS urlintro eprinturl eprintpr...

work page

[2] [2]

write newline

" write newline "" before.all 'output.state := FUNCTION n.dashify 't := "" t empty not t #1 #1 substring "-" = t #1 #2 substring "--" = not "--" * t #2 global.max substring 't := t #1 #1 substring "-" = "-" * t #2 global.max substring 't := while if t #1 #1 substring * t #2 global.max substring 't := if while FUNCTION word.in bbl.in capitalize " " * FUNCT...

work page

[3] [3]

Dzmitry Bahdanau, Kyunghyunand Cho, and Yoshua Bengio. 2015. Neural Machine Translation by Jointly Learning to Align and Translate . In Proceedings of the 3rd International Conference on Learning Representations (ICLR)

work page 2015

[4] [4]

Dasha Bogdanova and Jennifer Foster. 2016. This is how we do it: Answer Reranking for Open-Domain How Questions with Paragraph Vectors and Minimal Feature Engineering . In Proceedings of the 15th Annual Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (NAACL-HLT), pages 1290--1295

work page 2016

[5] [5]

Kyunghyun Cho, Bart van Merrienboer, Caglar Gulcehre, Dzmitry Bahdanau, Fethi Bougares, Holger Schwenk, and Yoshua Bengio. 2014. Learning Phrase Representations Using RNN Encoder-Decoder for Statistical Machine Translation . In Proceedings of the 2014 Conference on Empirical Methods in Natural Language Processing (EMNLP)

work page 2014

[6] [6]

Junyoung Chung, Caglar Gulcehre, KyungHyun Cho, and Yoshua Bengio. 2014. Empirical Evaluation of Gated Recurrent Neural Networks on Sequence Modeling . In Proceedings of the 28th Conference Neural Information Processing Systems, Deep Learning and Representation Learning Workshop (NIPS)

work page 2014

[7] [7]

George Doddington. 2002. Automatic Evaluation of Machine Translation Quality Using N-gram Co-occurrence Statistics . In Proceedings of the 2nd International Conference on Human Language Technology Research (HLT), pages 138--145

work page 2002

[8] [8]

Motoyasu Fujita, Rafal Rzepka, and Kenji Araki. 2011. Evaluation of Utterances Based on Causal Knowledge Retrieved from Blogs . In Proceedings of the 14th IASTED International Conference Artificial Intelligence and Soft Computing (ASC), pages 294--299

work page 2011

[9] [9]

Forgues Gabriel, Joelle Pineau, Jean-Marie Larchev \^ e que, and R \' e al Tremblay. 2014. Bootstrapping Dialog Systems with Word Embeddings

work page 2014

[10] [10]

Peter Jansen, Mihai Surdeanu, and Peter Clark. 2014. Discourse Complements Lexical Semantics for Non-factoid Answer Reranking . In Proceedings of the 52nd Annual Meeting of the Association for Computational Linguistics (ACL), pages 977--986

work page 2014

[11] [11]

Daisuke Kawahara and Sadao Kurohashi. 2006. A Fully-Lexicalized Probabilistic Model for Japanese Syntactic and Case Structure Analysis . In Proceedings of Human Language Technology Conference - North American Chapter of the Association for Computational Linguistics Annual Meeting (HLT-NAACL), pages 176--183

work page 2006

[12] [12]

Kingma and Jimmy Ba

Diederik P. Kingma and Jimmy Ba. 2015. Adam: A Method for Stochastic Optimization . In Proceedings of the 3rd International Conference on Learning Representations (ICLR)

work page 2015

[13] [13]

Taku Kudo and John Richardson. 2018. SentencePiece: A Simple and Language Independent Subword Tokenizer and Detokenizer for Neural Text Processing . In Proceedings of the 2018 Conference on Empirical Methods in Natural Language Processing (EMNLP)

work page 2018

[14] [14]

Jiwei Li, Michel Galley, Chris Brockett, Jianfeng Gao, and Bill Dolan. 2016. A Diversity-Promoting Objective Function for Neural Conversation Models . In Proceedings of the 15th Annual Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (NAACL-HLT), pages 110--119

work page 2016

[15] [15]

Serban, Michael Noseworthy, Laurent Charlin, and Joelle Pineau

Chia-Wei Liu, Ryan Lowe, Iulian V. Serban, Michael Noseworthy, Laurent Charlin, and Joelle Pineau. 2016. How NOT to Evaluate Your Dialogue System: An Empirical Study of Unsupervised Evaluation Metrics for Dialogue Response Generation . In Proceedings of the 2016 Conference on Empirical Methods in Natural Language Processing (EMNLP)

work page 2016

[16] [16]

Minh-Thang Luong, Hieu Pham, and Christopher D. Manning. 2015. Effective Approaches to Attention-Based Neural Machine Translation . In Proceedings of the 2015 Conference on Empirical Methods in Natural Language Processing (EMNLP)

work page 2015

[17] [17]

Wolfgang Macherey, Maxim Krikun, Yuan Cao, Qin Gao, Klaus Macherey, Jeff Klingner, Apurva Shah, Melvin Johnson, Xiaobing Liu, ukasz Kaiser, Stephan Gouws, Yoshikiyo Kato, Taku Kudo, Hideto Kazawa, Keith Stevens, George Kurian, Nishant Patil, Wei Wang, Cliff Young, Jason Smith, Jason Riesa, Alex Rudnick, Oriol Vinyals, Greg Corrado, Macduff Hughes, and Jef...

work page internal anchor Pith review Pith/arXiv arXiv 2016

[18] [18]

Tomas Mikolov, Kai Chen, Greg Corrado, and Jeffrey Deany. 2013 a . Efficient Estimation of Word Representations in Vector Space . In Proceedings of the 1st International Conference on Learning Representations (ICLR)

work page 2013

[19] [19]

Tomas Mikolov, Ilya Sutskever, Kai Chen, Greg Corrado, and Jeffrey Dean. 2013 b . Distributed Representations of Words and Phrases and Their Compositionality . In Proceedings of the 26th International Conference on Neural Information Processing Systems (NIPS), volume 2, pages 3111--3119

work page 2013

[20] [20]

Tomas Mikolov, Wen-tau Yih, and Geoffrey Zweig. 2013 c . Linguistic Regularities in Continuous Space Word Representations . In Proceedings of the 12th Annual Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (NAACL-HLT), pages 746--751

work page 2013

[21] [21]

Ryo Nakamura, Katsuhito Sudoh, Koichiro Yoshino, and Satoshi Nakamura. 2019. Another Diversity-Promoting Objective Function for Neural Dialogue Generation . In Proceedings of the 33rd Association for the Advancement of Artificial Intelligence Conference on Artificial Intelligence, Workshop on Reasoning and Learning for Human-Machine Dialogues (DEEP-DIAL 2...

work page 2019

[22] [22]

David Newman, Jey Han Lau, Karl Grieser, and Timothy Baldwin. 2010. Automatic Evaluation of Topic Coherence . In Proceedings of the 11th Annual Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (NAACL-HLT), pages 100--108

work page 2010

[23] [23]

Jong-Hoon Oh, Kentaro Torisawa, Chikara Hashimoto, Ryu Iida, Masahiro Tanaka, and Julien Kloetzer. 2016. A Semi-supervised Learning Approach to Why-Question Answering . In Proceedings of the 30th Association for the Advancement of Artificial Intelligence Conference on Artificial Intelligence (AAAI), pages 3022--3029

work page 2016

[24] [24]

Jong-Hoon Oh, Kentaro Torisawa, Chikara Hashimoto, Motoki Sano, Stijn De Saeger, and Kiyonori Ohtake. 2013. Why-Question Answering Using Intra- and Inter-Sentential Causal Relations . In Proceedings of the 51st Annual Meeting of the Association for Computational Linguistics (ACL), pages 1733--1743

work page 2013

[25] [25]

Jong-Hoon Oh, Kentaro Torisawa, Canasai Kruengkrai, Ryu Iida, and Julien Kloetzer. 2017. Multi-Column Convolutional Neural Networks with Causality-Attention for Why-Question Answering . In Proceedings of the 10th Association for Computing Machinery International Conference on Web Search and Data Mining (WSDM), pages 415--424

work page 2017

[26] [26]

Junki Ohmura and Maxine Eskenazi. 2018. Context-Aware Dialog Re-ranking for Task-Oriented Dialog Systems . In Proceedings of IEEE Spoken Language Technology Workshop (SLT)

work page 2018

[27] [27]

Kishore Papineni, Salim Roukos, Todd Ward, and Wei-Jing Zhu. 2002. BLEU: a Method for Automatic Evaluation of Machine Translation . In Proceedings of the 40th Annual Meeting of the Association for Computational Linguistics (ACL), pages 311--318

work page 2002

[28] [28]

Ashwin Ram, Rohit Prasad, Chandra Khatri, Anu Venkatesh, Raefer Gabriel, Qing Liu, Jeff Nunn, Behnam Hedayatnia, Ming Cheng, Ashish Nagar, Eric King, Kate Bland, Amanda Wartick, Yi Pan, Han Song, Sk Jayadevan, Gene Hwang, and Art Pettigrue. 2018. Conversational AI: The Science Behind the Alexa Prize . In arXiv:1801.03604

work page internal anchor Pith review Pith/arXiv arXiv 2018

[29] [29]

Ryohei Sasano and Sadao Kurohashi. 2011. A Discriminative Approach to Japanese Zero Anaphora Resolution with Large-Scale Lexicalized Case Frames . In Proceedings of the 5th International Joint Conference on Natural Language Processing (IJCNLP), pages 758--766

work page 2011

[30] [30]

Serban, Alessandro Sordoni, Yoshua Bengio, Aaron Courville, and Joelle Pineau

Iulian V. Serban, Alessandro Sordoni, Yoshua Bengio, Aaron Courville, and Joelle Pineau. 2016. Building End-To-End Dialogue Systems Using Generative Hierarchical Neural Network Models . In Proceedings of the 30th Association for the Advancement of Artificial Intelligence Conference on Artificial Intelligence (AAAI)

work page 2016

[31] [31]

Tomohide Shibata, Shotaro Kohama, and Sadao Kurohashi. 2014. A Large Scale Database of Strongly-Related Events in Japanese . In Proceedings of the 9th International Conference on Language Resources and Evalu ation (LREC)

work page 2014

[32] [32]

Tomohide Shibata and Sadao Kurohashi. 2011. Acquiring Strongly-Related Events Using Predicate-Argument Co-occurring Statist ics and Case Frames . In Proceedings of the 5th International Joint Conference on Natural Language Proce ssing (IJCNLP), pages 1028--1036

work page 2011

[33] [33]

Simonsen, and Jian-Yun Nie

Alessandro Sordoni, Yoshua Bengio, Hossein Vahabi, Christina Lioma, Jakob G. Simonsen, and Jian-Yun Nie. 2015. A Hierarchical Recurrent Encoder-Decoder For Generative Context-Aware Query Suggestion . In Proceedings of the 24th Association for Computing Machinery International Conference on Information Knowledge and Management (ACM)

work page 2015

[34] [34]

Oriol Vinyals and Quoc V. Le. 2015. A Neural Conversational Model . In Proceedings of the 32nd International Conference on Machine Learning, Deep Learning Workshop (ICML)

work page 2015

[35] [35]

Noah Weber, Niranjan Balasubramanian, and Nathanael Chambers. 2018. Event Representations with Tensor-Based Compositions . In Proceedings of the 32nd Association for the Advancement of Artificial Intelligence Conference on Artificial Intelligence (AAAI)

work page 2018