Filters








9,780 Hits in 3.8 sec

Generating Persona-Consistent Dialogue Responses Using Deep Reinforcement Learning [article]

Mohsen Mesgar, Edwin Simpson, Yue Wang, Iryna Gurevych
2020 arXiv   pre-print
We propose a novel approach to train transformer-based dialogue agents using actor-critic reinforcement learning.  ...  We define a new reward function to assess generated responses in terms of persona consistency, topic consistency, and fluency.  ...  We propose a novel approach to train dialogue agents with deep reinforcement learning (DeepRL).  ... 
arXiv:2005.00036v1 fatcat:downot5rifaxvnw23hmnv35fbe

Generating Persona Consistent Dialogues by Exploiting Natural Language Inference [article]

Haoyu Song, Wei-Nan Zhang, Jingwen Hu, Ting Liu
2019 arXiv   pre-print
the persona-consistency of generated responses.  ...  Different from existing work that re-ranks the retrieved responses through an NLI model, we cast the task as a reinforcement learning problem and propose to exploit the NLI signals from response-persona  ...  Reinforcement Learning We formalize the persona consistent dialogue generation problem as a reinforcement learning task.  ... 
arXiv:1911.05889v3 fatcat:rl3ta7j7wnchtgvbvpt4bej3ju

Generating Persona Consistent Dialogues by Exploiting Natural Language Inference

Haoyu Song, Wei-Nan Zhang, Jingwen Hu, Ting Liu
2020 PROCEEDINGS OF THE THIRTIETH AAAI CONFERENCE ON ARTIFICIAL INTELLIGENCE AND THE TWENTY-EIGHTH INNOVATIVE APPLICATIONS OF ARTIFICIAL INTELLIGENCE CONFERENCE  
the persona-consistency of generated responses.  ...  Different from existing work that re-ranks the retrieved responses through an NLI model, we cast the task as a reinforcement learning problem and propose to exploit the NLI signals from response-persona  ...  Reinforcement Learning We formalize the persona consistent dialogue generation problem as a reinforcement learning task.  ... 
doi:10.1609/aaai.v34i05.6417 fatcat:46hhzqra5bavth7ct6lpz3g4gm

Deep Active Learning for Dialogue Generation [article]

Nabiha Asghar, Pascal Poupart, Xin Jiang, Hang Li
2017 arXiv   pre-print
Experiments show that our model inherently promotes the generation of semantically relevant and interesting responses, and can be used to train agents with customized personas, moods and conversational  ...  We propose an online, end-to-end, neural generative conversational model for open-domain dialogue.  ...  Researchers are now exploring Deep Reinforcement Learning (DRL) to address the hard problems of NLU and NLG in dialogue generation.  ... 
arXiv:1612.03929v5 fatcat:b26h4odcvnb6bffoozvxgm2evq

Deep Active Learning for Dialogue Generation

Nabiha Asghar, Pascal Poupart, Xin Jiang, Hang Li
2017 Proceedings of the 6th Joint Conference on Lexical and Computational Semantics (*SEM 2017)  
Experiments show that our model inherently promotes the generation of semantically relevant and interesting responses, and can be used to train agents with customized personas, moods and conversational  ...  We propose an online, end-to-end, neural generative conversational model for opendomain dialogue.  ...  Researchers are now exploring Deep Reinforcement Learning (DRL) to address the hard problems of NLU and NLG in dialogue generation.  ... 
doi:10.18653/v1/s17-1008 dblp:conf/starsem/AsgharPJL17 fatcat:pfimq5yzf5bjnibxnxscqpkdsm

Recent Progress in Conversational AI [article]

Zijun Xue, Ruirui Li, Mingda Li
2022 arXiv   pre-print
We will provide a brief review of the recent progress in the Conversational AI, including the commonly adopted techniques, notable works, famous competitions from academia and industry and widely used  ...  In [49] , the author proposes a hierarchical deep reinforcement learning framework to solve the composite task-completion dialogue policy learning problem.  ...  In [53] [40] , the author proposed an acceleration technique by using the Thompson sampling to improve the efficiency of the deep reinforcement learning(DQN).  ... 
arXiv:2204.09719v1 fatcat:lf7g4enbsfc2rona3ki3cuwsua

Recent Advances in Deep Learning Based Dialogue Systems: A Systematic Survey [article]

Jinjie Ni, Tom Young, Vlad Pandelea, Fuzhao Xue, Erik Cambria
2022 arXiv   pre-print
In this survey, we mainly focus on the deep learning based dialogue systems.  ...  To the best of our knowledge, this survey is the most comprehensive and up-to-date one at present for deep learning based dialogue systems, extensively covering the popular techniques.  ...  RL for knowledge grounded dialogue systems Some systems use reinforcement learning to select from outside information like persona, document, knowledge graph, etc., and generate responses accordingly.  ... 
arXiv:2105.04387v5 fatcat:yd3gqg45rjgzxbiwfdlcvf3pye

The Rapidly Changing Landscape of Conversational Agents [article]

Vinayak Mathur, Arpit Singh
2018 arXiv   pre-print
We look at statistical, neural, generative adversarial network based and reinforcement learning based approaches and how they evolved.  ...  persona etc.  ...  Actor-Critic Algorithm Deep reinforcement learning (RL) methods have significant potential for dialogue policy optimisation.  ... 
arXiv:1803.08419v2 fatcat:5huy2e6tmbexlfoampcdy6zhw4

Neural Approaches to Conversational AI

Jianfeng Gao, Michel Galley, Lihong Li
2018 Proceedings of ACL 2018, Tutorial Abstracts  
We group conversational systems into three categories: (1) question answering agents, (2) taskoriented dialogue agents, and (3) social bots.  ...  a review of state-of-the-art neural approaches, draw the connection between neural approaches and traditional symbolic approaches, and discuss the progress we have made and challenges we are facing, using  ...  ., 2016b) presented a persona-based model to address the issue of speaker consistency in neural response generation.  ... 
doi:10.18653/v1/p18-5002 dblp:conf/acl/GaoGL18 fatcat:7llxwuntafh4fcjj4ia3tm642a

A Task-oriented Chatbot Based on LSTM and Reinforcement Learning

Yu-Ling Hsueh, Tai-Liang Chou
2022 ACM Transactions on Asian and Low-Resource Language Information Processing  
Thanks to the advancements in deep learning, chatbots are widely used in messaging applications. Undoubtedly, a chatbot is a new way of interaction between humans and machines.  ...  Hence, training a chatbot model that uses low-resource conversational data to generate more diverse dialogues is desirable.  ...  The deep learning system with reinforcement learning generates a response given to an input of a user. The detailed description of the related topics is listed as follows.  ... 
doi:10.1145/3529649 fatcat:swzdgoy6ubahbmdd6mxim4c7mu

Deep Reinforcement Learning for Chatbots Using Clustered Actions and Human-Likeness Rewards

Heriberto Cuayahuitl, Donghyeon Lee, Seonghan Ryu, Sungja Choi, Inchul Hwang, Jihie Kim
2019 2019 International Joint Conference on Neural Networks (IJCNN)  
We train Deep Reinforcement Learning (DRL) agents using chitchat data in raw text-without any manual annotations. Experimental results using different splits of training data report the following.  ...  Training chatbots using the reinforcement learning paradigm is challenging due to high-dimensional states, infinite action spaces and the difficulty in specifying the reward function.  ...  We present a novel approach based on the reinforcement learning [2] , unsupervised learning [3] and deep learning [4] paradigms.  ... 
doi:10.1109/ijcnn.2019.8852376 dblp:conf/ijcnn/CuayahuitlLRCHK19 fatcat:dvtylbfegrc2fey55hcguc3hti

State of the Art of User Simulation approaches for conversational information retrieval [article]

Pierre Erbacher and Laure Soulier and Ludovic Denoyer
2022 arXiv   pre-print
Reinforcement learning has emerged as a paradigm particularly suited to optimize sequential decision making in many domains and has recently appeared in IR.  ...  However, training these systems by reinforcement learning on users is not feasible. One solution is to train IR systems on user simulations that model the behavior of real users.  ...  Some works propose to use US with attributes and characteristics (persona) [51, 52] to obtain both diversity and consistency in requests and dialogues. Li et al.  ... 
arXiv:2201.03435v1 fatcat:4yiydzysqzddvaju3jjoph7sdu

Ensemble-Based Deep Reinforcement Learning for Chatbots

Heriberto Cuayáhuitl, Donghyeon Lee, Seonghan Ryu, Yongjin Cho, Sungja Choi, Satish Indurthi, Seunghak Yu, Hyungtak Choi, Inchul Hwang, Jihie Kim
2019 Neurocomputing  
Deep Reinforcement Learning (DRL) is promising for addressing this challenge, but its successful application remains an open question.  ...  In addition to evaluations using held-out data, our results are further supported by a human evaluation that rated dialogues in terms of fluency, engagingness and consistency -- which revealed that our  ...  A deep reinforcement learning agent approximates Q * using a multi-layer neural network [31] .  ... 
doi:10.1016/j.neucom.2019.08.007 fatcat:whpcdh5zxzdnlmwvrm6lru736u

Deep Reinforcement Learning for Chatbots Using Clustered Actions and Human-Likeness Rewards [article]

Heriberto Cuayáhuitl, Donghyeon Lee, Seonghan Ryu, Sungja Choi, Inchul Hwang, Jihie Kim
2019 arXiv   pre-print
We train Deep Reinforcement Learning (DRL) agents using chitchat data in raw text---without any manual annotations. Experimental results using different splits of training data report the following.  ...  Training chatbots using the reinforcement learning paradigm is challenging due to high-dimensional states, infinite action spaces and the difficulty in specifying the reward function.  ...  High-level architecture of the proposed deep reinforcement learning approach for chatbots-see text for details provided labels), a Deep Reinforcement Learning (DRL) agent takes the role of one of the two  ... 
arXiv:1908.10331v1 fatcat:jtgyrwly4jfaxfeb7u5mbbxz7a

Deep Reinforcement Learning for Conversational AI [article]

Mahipal Jadeja, Neelanshi Varia, Agam Shah
2017 arXiv   pre-print
It is possible to scale deep reinforcement learning with the use of deep learning and do amazing tasks such as use of pixels in playing video games.  ...  Various conversational models which are based on deep reinforcement learning (as well as deep learning) are also discussed.  ...  Challenge 3: Coherent dialogue design e agent should generate consistent response while generating answer for semantically identical input.  ... 
arXiv:1709.05067v1 fatcat:ogg6iej4jjdflp2wcvpwxmfegi
« Previous Showing results 1 — 15 out of 9,780 results