Simulated Chats for Building Dialog Systems: Learning to Generate Conversations from Instructions

Popular dialog data sets such as MultiWOZ (Budzianowski et al., 2018) are created by providing crowd workers an instruction, expressed in natural language, that describes the task to be accomplished. Crowd workers play the role of a user and an agent to generate dialogs to accomplish tasks involving booking restaurant tables, calling a taxi etc. In this paper, we present a data creation strategy that uses the pre-trained language model, GPT2 (Radford et al., 2018), to simulate the interaction between crowd workers by creating a user bot and an agent bot. We train the simulators using a smaller percentage of actual crowd-generated conversations and their corresponding instructions. We demonstrate that by using the simulated data, we achieve significant improvements in low-resource settings on two publicly available datasets MultiWOZ dataset (Budzianowski et al., 2018) and the Persona chat dataset (Zhang et al., 2018a).

[1]  Mark Chen,et al.  Language Models are Few-Shot Learners , 2020, NeurIPS.

[2]  Zhijian Ou,et al.  Paraphrase Augmented Task-Oriented Dialog Generation , 2020, ACL.

[3]  Maryam Fazel-Zarandi,et al.  Data Augmentation for Training Dialog Models Robust to Speech Recognition Errors , 2020, NLP4CONVAI.

[4]  Jason Weston,et al.  Dialogue Natural Language Inference , 2018, ACL.

[5]  Joelle Pineau,et al.  The Second Conversational Intelligence Challenge (ConvAI2) , 2019, The NeurIPS '18 Competition.

[6]  Colin Raffel,et al.  Exploring the Limits of Transfer Learning with a Unified Text-to-Text Transformer , 2019, J. Mach. Learn. Res..

[7]  Lijun Wu,et al.  Learning to Teach with Dynamic Loss Functions , 2018, NeurIPS.

[8]  Arman Cohan,et al.  Longformer: The Long-Document Transformer , 2020, ArXiv.

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

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

[11]  Nir Ailon,et al.  Deep Metric Learning Using Triplet Network , 2014, SIMBAD.

[12]  Christof Monz,et al.  Optimizing Transformer for Low-Resource Neural Machine Translation , 2020, COLING.

[13]  Angeliki Metallinou,et al.  Controlled Text Generation for Data Augmentation in Intelligent Artificial Agents , 2019, EMNLP.

[14]  Sang-goo Lee,et al.  Data Augmentation for Spoken Language Understanding via Joint Variational Generation , 2018, AAAI.

[15]  Dilek Z. Hakkani-Tür,et al.  Dialog Simulation with Realistic Variations for Training Goal-Oriented Conversational Systems , 2020, ArXiv.

[16]  Pascale Fung,et al.  MinTL: Minimalist Transfer Learning for Task-Oriented Dialogue Systems , 2020, EMNLP.

[17]  Zhou Yu,et al.  A Student-Teacher Architecture for Dialog Domain Adaptation under the Meta-Learning Setting , 2021, AAAI.

[18]  Geoffrey E. Hinton,et al.  Distilling the Knowledge in a Neural Network , 2015, ArXiv.

[19]  Zhijian Ou,et al.  Task-Oriented Dialog Systems that Consider Multiple Appropriate Responses under the Same Context , 2019, AAAI.

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

[21]  Jason Baldridge,et al.  PAWS: Paraphrase Adversaries from Word Scrambling , 2019, NAACL.

[22]  Zi-Yi Dou,et al.  Investigating Meta-Learning Algorithms for Low-Resource Natural Language Understanding Tasks , 2019, EMNLP.

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

[24]  Baolin Peng,et al.  Soloist: Building Task Bots at Scale with Transfer Learning and Machine Teaching , 2021, Transactions of the Association for Computational Linguistics.

[25]  Gökhan Tür,et al.  Building a Conversational Agent Overnight with Dialogue Self-Play , 2018, ArXiv.

[26]  Alan W. Black,et al.  Data Augmentation for Neural Online Chats Response Selection , 2018, SCAI@EMNLP.

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

[28]  Quoc V. Le,et al.  Self-Training With Noisy Student Improves ImageNet Classification , 2019, 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR).

[29]  Samy Bengio,et al.  Large Scale Online Learning of Image Similarity Through Ranking , 2009, J. Mach. Learn. Res..

[30]  Maxine Eskénazi,et al.  Structured Fusion Networks for Dialog , 2019, SIGdial.

[31]  Thomas Wolf,et al.  HuggingFace's Transformers: State-of-the-art Natural Language Processing , 2019, ArXiv.

[32]  Ateret Anaby-Tavor,et al.  Do Not Have Enough Data? Deep Learning to the Rescue! , 2020, AAAI.

[33]  Jian Zhang,et al.  SQuAD: 100,000+ Questions for Machine Comprehension of Text , 2016, EMNLP.

[34]  Wanxiang Che,et al.  A Corpus-free State2Seq User Simulator for Task-oriented Dialogue , 2019, CCL.

[35]  Gökhan Tür,et al.  Collaborative Multi-Agent Dialogue Model Training Via Reinforcement Learning , 2019, SIGdial.

[36]  Eunah Cho,et al.  Data Augmentation using Pre-trained Transformer Models , 2020, LIFELONGNLP.