Alternating Recurrent Dialog Model with Large-scale Pre-trained Language Models

Existing dialog system models require extensive human annotations and are difficult to generalize to different tasks. The recent success of large pre-trained language models such as BERT and GPT-2 (Devlin et al., 2019; Radford et al., 2019) have suggested the effectiveness of incorporating language priors in down-stream NLP tasks. However, how much pre-trained language models can help dialog response generation is still under exploration. In this paper, we propose a simple, general, and effective framework: Alternating Roles Dialog Model (ARDM). ARDM models each speaker separately and takes advantage of the large pre-trained language model. It requires no supervision from human annotations such as belief states or dialog acts to achieve effective conversations. ARDM outperforms or is on par with state-of-the-art methods on two popular task-oriented dialog datasets: CamRest676 and MultiWOZ. Moreover, we can generalize ARDM to more challenging, non-collaborative tasks such as persuasion. In persuasion tasks, ARDM is capable of generating human-like responses to persuade people to donate to a charity.

[1]  Frank Hutter,et al.  Decoupled Weight Decay Regularization , 2017, ICLR.

[2]  Kenneth Heafield,et al.  Proceedings of the 54th Annual Meeting of the Association for Computational Linguistics, ACL 2016, August 7-12, 2016, Berlin, Germany, Volume 1: Long Papers , 2016, Annual Meeting of the Association for Computational Linguistics.

[3]  Thomas Wolf,et al.  TransferTransfo: A Transfer Learning Approach for Neural Network Based Conversational Agents , 2019, ArXiv.

[4]  Rico Sennrich,et al.  Neural Machine Translation of Rare Words with Subword Units , 2015, ACL.

[5]  Ta-Chung Chi,et al.  Speaker Role Contextual Modeling for Language Understanding and Dialogue Policy Learning , 2017, IJCNLP.

[6]  Ilya Sutskever,et al.  Language Models are Unsupervised Multitask Learners , 2019 .

[7]  Maxine Eskénazi,et al.  Rethinking Action Spaces for Reinforcement Learning in End-to-end Dialog Agents with Latent Variable Models , 2019, NAACL.

[8]  Demis Hassabis,et al.  Mastering Chess and Shogi by Self-Play with a General Reinforcement Learning Algorithm , 2017, ArXiv.

[9]  Alec Radford,et al.  Improving Language Understanding by Generative Pre-Training , 2018 .

[10]  Boi Faltings,et al.  Personalization in Goal-Oriented Dialog , 2017, ArXiv.

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

[12]  Christopher D. Manning,et al.  A Copy-Augmented Sequence-to-Sequence Architecture Gives Good Performance on Task-Oriented Dialogue , 2017, EACL.

[13]  Jason Weston,et al.  Personalizing Dialogue Agents: I have a dog, do you have pets too? , 2018, ACL.

[14]  Yiming Yang,et al.  Transformer-XL: Attentive Language Models beyond a Fixed-Length Context , 2019, ACL.

[15]  Jianfeng Gao,et al.  A Persona-Based Neural Conversation Model , 2016, ACL.

[16]  Yun-Nung Chen,et al.  How Time Matters: Learning Time-Decay Attention for Contextual Spoken Language Understanding in Dialogues , 2018, NAACL.

[17]  Lukasz Kaiser,et al.  Attention is All you Need , 2017, NIPS.

[18]  Ta-Chung Chi,et al.  Dynamic time-aware attention to speaker roles and contexts for spoken language understanding , 2017, 2017 IEEE Automatic Speech Recognition and Understanding Workshop (ASRU).

[19]  Tsung-Hsien Wen,et al.  Latent Intention Dialogue Models , 2017, ICML.

[20]  Stefan Ultes,et al.  MultiWOZ - A Large-Scale Multi-Domain Wizard-of-Oz Dataset for Task-Oriented Dialogue Modelling , 2018, EMNLP.

[21]  Zhou Yu,et al.  Persuasion for Good: Towards a Personalized Persuasive Dialogue System for Social Good , 2019, ACL.

[22]  Tatsuya Kawahara,et al.  Effective Incorporation of Speaker Information in Utterance Encoding in Dialog , 2019, ArXiv.

[23]  Yejin Choi,et al.  The Curious Case of Neural Text Degeneration , 2019, ICLR.

[24]  Ivan Vulić,et al.  Hello, It’s GPT-2 - How Can I Help You? Towards the Use of Pretrained Language Models for Task-Oriented Dialogue Systems , 2019, EMNLP.

[25]  Min-Yen Kan,et al.  Sequicity: Simplifying Task-oriented Dialogue Systems with Single Sequence-to-Sequence Architectures , 2018, ACL.

[26]  Wenhu Chen,et al.  Semantically Conditioned Dialog Response Generation via Hierarchical Disentangled Self-Attention , 2019, ACL.

[27]  Matthew Henderson,et al.  The Second Dialog State Tracking Challenge , 2014, SIGDIAL Conference.

[28]  Christopher D. Manning,et al.  Key-Value Retrieval Networks for Task-Oriented Dialogue , 2017, SIGDIAL Conference.

[29]  Yejin Choi,et al.  SWAG: A Large-Scale Adversarial Dataset for Grounded Commonsense Inference , 2018, EMNLP.