19,912 Hits in 4.6 sec

Multi-Referenced Training for Dialogue Response Generation [article]

Tianyu Zhao, Tatsuya Kawahara
2020 arXiv   pre-print
Then we explore approaches to multi-referenced training in two aspects.  ...  In open-domain dialogue response generation, a dialogue context can be continued with diverse responses, and the dialogue models should capture such one-to-many relations.  ...  In this paper, we will investigate why single-referenced training harms our dialogue models and how to apply multi-referenced training. Why Multi-Referenced Training Matters?  ... 
arXiv:2009.07117v2 fatcat:zru3p5kwwnaxtl5gtsuii53a2a

Better Automatic Evaluation of Open-Domain Dialogue Systems with Contextualized Embeddings [article]

Sarik Ghazarian, Johnny Tian-Zheng Wei, Aram Galstyan, Nanyun Peng
2019 arXiv   pre-print
A recent work proposed Referenced metric and Unreferenced metric Blended Evaluation Routine (RUBER) to combine a learning-based metric, which predicts relatedness between a generated response and a given  ...  Traditional reference-based metrics such as BLEU are ineffective because there could be many valid responses for a given context that share no common words with reference responses.  ...  Acknowledgments We thank the anonymous reviewers for their constructive feedback, as well as the members of the PLUS lab for their useful discussion and feedback.  ... 
arXiv:1904.10635v1 fatcat:n3qzpwd3dzeejiau2qgy23zzfy

Designing Precise and Robust Dialogue Response Evaluators [article]

Tianyu Zhao, Divesh Lala, Tatsuya Kawahara
2020 arXiv   pre-print
Automatic dialogue response evaluator has been proposed as an alternative to automated metrics and human evaluation.  ...  Experimental results demonstrate that the proposed evaluator achieves a strong correlation (> 0.6) with human judgement and generalizes robustly to diverse responses and corpora.  ...  Acknowledgments The authors would like to thank Shinsuke Mori from Kyoto University, Wei Wu from Microsoft, Graham Neubig from CMU, and the anonymous reviewers for their constructive comments.  ... 
arXiv:2004.04908v2 fatcat:hxacualaurhyjpbw5z2qbeeagi

Investigating Evaluation of Open-Domain Dialogue Systems With Human Generated Multiple References [article]

Prakhar Gupta, Shikib Mehri, Tiancheng Zhao, Amy Pavel, Maxine Eskenazi, Jeffrey P. Bigham
2019 arXiv   pre-print
One alternative is to collect human annotations for evaluation, which can be expensive and time consuming.  ...  To demonstrate the effectiveness of multi-reference evaluation, we augment the test set of DailyDialog with multiple references.  ...  Referenced Diversity A multi-reference test set also allows referenced diversity evaluation.  ... 
arXiv:1907.10568v2 fatcat:tzxkifuhzbdb5noypdoq52fqqm

Learning an Unreferenced Metric for Online Dialogue Evaluation [article]

Koustuv Sinha, Prasanna Parthasarathi, Jasmine Wang, Ryan Lowe, William L. Hamilton, Joelle Pineau
2020 arXiv   pre-print
There have been recent efforts to develop automatic dialogue evaluation metrics, but most of them do not generalize to unseen datasets and/or need a human-generated reference response during inference,  ...  We show that our model achieves higher correlation with human annotations in an online setting, while not requiring true responses for comparison during inference.  ...  We would like to thank Shagun Sodhani and Alborz Geramifard for helpful feedback on the manuscript.  ... 
arXiv:2005.00583v1 fatcat:o2aw5ywxcffbdehq7p3eajirhq

Incorporating Interlocutor-Aware Context into Response Generation on Multi-Party Chatbots [article]

Cao Liu, Kang Liu, Shizhu He, Zaiqing Nie, Jun Zhao
2019 arXiv   pre-print
Conventional chatbots focus on two-party response generation, which simplifies the real dialogue scene.  ...  In this paper, we strive toward a novel task of Response Generation on Multi-Party Chatbot (RGMPC), where the generated responses heavily rely on the interlocutors' roles (e.g., speaker and addressee)  ...  Finally, the interlocutor prediction loss is added to the response generation loss for training.  ... 
arXiv:1910.13106v1 fatcat:lezhqdl4evgurljpt44ugqbufq

MultiWOZ 2.3: A multi-domain task-oriented dialogue dataset enhanced with annotation corrections and co-reference annotation [article]

Ting Han, Ximing Liu, Ryuichi Takanobu, Yixin Lian, Chongxuan Huang, Dazhen Wan, Wei Peng, Minlie Huang
2021 arXiv   pre-print
To ensure consistency between dialogue acts and dialogue states, we implement co-reference features and unify annotations of dialogue acts and dialogue states.  ...  In this paper, we introduce MultiWOZ 2.3, in which we differentiate incorrect annotations in dialogue acts from dialogue states, identifying a lack of co-reference when publishing the updated dataset.  ...  " slots for taxi/train domains.  ... 
arXiv:2010.05594v3 fatcat:t2o46xv5ybc25ilow2bo3y3sui

Commonsense-Focused Dialogues for Response Generation: An Empirical Study [article]

Pei Zhou, Karthik Gopalakrishnan, Behnam Hedayatnia, Seokhwan Kim, Jay Pujara, Xiang Ren, Yang Liu, Dilek Hakkani-Tur
2021 arXiv   pre-print
We evaluate response generation models trained using these datasets and find that models trained on both extracted and our collected data produce responses that consistently exhibit more commonsense than  ...  Moreover, existing dialogue datasets do not explicitly focus on exhibiting commonsense as a facet. In this paper, we present an empirical study of commonsense in dialogue response generation.  ...  Aiming to bridge the gap in commonsense for dialogue response generation, we collect a large-scale multi-turn open-domain dialogue dataset that is focused on commonsense knowledge.  ... 
arXiv:2109.06427v2 fatcat:c3ktn55vyzhhhkfskn4owddu7m

Multi-Task Pre-Training for Plug-and-Play Task-Oriented Dialogue System [article]

Yixuan Su, Lei Shu, Elman Mansimov, Arshit Gupta, Deng Cai, Yi-An Lai, Yi Zhang
2022 arXiv   pre-print
In addition, we introduce a new dialogue multi-task pre-training strategy that allows the model to learn the primary TOD task completion skills from heterogeneous dialog corpora.  ...  In this study, we present PPTOD, a unified plug-and-play model for task-oriented dialogue.  ...  for generating dialogue responses.  ... 
arXiv:2109.14739v2 fatcat:dsy25xhxlbbk3f2qccoluwtpiy

Integrating User and Agent Models: A Deep Task-Oriented Dialogue System [article]

Weiyan Wang, Yuxiang WU, Yu Zhang, Zhongqi Lu, Kaixiang Mo, Qiang Yang
2017 arXiv   pre-print
Then the built user model is used as a leverage to train the agent model by deep reinforcement learning.  ...  Task-oriented dialogue systems can efficiently serve a large number of customers and relieve people from tedious works.  ...  Experiments on Multi-Turn Dialogue Generation In this section, we evaluate the proposed SAMIA framework to generate responses in multi-turn dialogues.  ... 
arXiv:1711.03697v1 fatcat:u67st24k6zdmvkzjyhqejphuci

Beyond Goldfish Memory: Long-Term Open-Domain Conversation [article]

Jing Xu, Arthur Szlam, Jason Weston
2021 arXiv   pre-print
Despite recent improvements in open-domain dialogue models, state of the art models are trained and evaluated on short conversations with little context.  ...  We show how existing models trained on existing datasets perform poorly in this long-term conversation setting in both automatic and human evaluations, and we study long-context models that can perform  ...  A memory-augmented generator that takes the dialogue context and access to the long-term memory, and then generates the next response.  ... 
arXiv:2107.07567v1 fatcat:wrabh7xfcba67n2nj6jfbqlesq

RiSAWOZ: A Large-Scale Multi-Domain Wizard-of-Oz Dataset with Rich Semantic Annotations for Task-Oriented Dialogue Modeling [article]

Jun Quan, Shian Zhang, Qian Cao, Zizhong Li, Deyi Xiong
2020 arXiv   pre-print
Both single- and multi-domain dialogues are constructed, accounting for 65% and 35%, respectively.  ...  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  ...  We would like to thank the anonymous reviewers for their insightful comments. The corresponding author is Deyi Xiong (  ... 
arXiv:2010.08738v1 fatcat:2qdeo3lcdbg6xl27y6q3sz7jl4

A Persona-based Multi-turn Conversation Model in an Adversarial Learning Framework [article]

Oluwatobi O. Olabiyi, Anish Khazane, Erik T. Mueller
2019 arXiv   pre-print
In this paper, we extend the persona-based sequence-to-sequence (Seq2Seq) neural network conversation model to multi-turn dialogue by modifying the state-of-the-art hredGAN architecture.  ...  The resulting persona hredGAN (phredGAN) shows better performance than both the existing persona-based Seq2Seq and hredGAN models when those external attributes are available in a multi-turn dialogue corpus  ...  Problem Formulation: The hredGAN [7] formulates multi-turn dialogue response generation as: given a dialogue history of sequence of utterances, X i = X 1 , X 2 , · · · , X i , where each utterance X  ... 
arXiv:1905.01998v1 fatcat:bluopydsazaeperznwjouplk5u

Low-Resource Adaptation of Open-Domain Generative Chatbots [article]

Greyson Gerhard-Young, Raviteja Anantha, Srinivas Chappidi, Björn Hoffmeister
2022 arXiv   pre-print
Additionally, we propose a generic framework that accounts for variety in question types, tracks reference throughout multi-turn conversations, and removes inconsistent and potentially toxic responses.  ...  We demonstrate that low parameter models can simultaneously retain their general knowledge conversational abilities while improving in a specific domain.  ...  The core response generation model was trained using the ParlAI 2 framework, a platform designed specifically for dialogue models.  ... 
arXiv:2108.06329v2 fatcat:ayloa42sozbpfbyeyuarroa6si

Pchatbot: A Large-Scale Dataset for Personalized Chatbot [article]

Hongjin Qian, Xiaohe Li, Hanxun Zhong, Yu Guo, Yueyuan Ma, Yutao Zhu, Zhanliang Liu, Zhicheng Dou, Ji-Rong Wen
2021 arXiv   pre-print
Different from existing datasets, Pchatbot provides anonymized user IDs and timestamps for both posts and responses.  ...  Besides, current dialogue datasets for personalized chatbot usually contain several persona sentences or attributes.  ...  (2) Multi-referenced dialogue. In natural language dialogue, a post can have multiple appropriate responses.  ... 
arXiv:2009.13284v3 fatcat:x47nvif75zae5oy53fd67shrxe
« Previous Showing results 1 — 15 out of 19,912 results