TransferTransfo: A Transfer Learning Approach for Neural Network Based Conversational Agents

Abstract

We introduce a new approach to generative data-driven dialogue systems (e.g. chatbots) called TransferTransfo which is a combination of a Transfer learning based training scheme and a high-capacity Transformer model. Fine-tuning is performed by using a multi-task objective which combines several unsupervised prediction tasks. The resulting fine-tuned model shows strong improvements over the current state-of-the-art end-to-end conversational models like memory augmented seq2seq and information-retrieval models. On the privately held PERSONA-CHAT dataset of the Conversational Intelligence Challenge 2, this approach obtains a new state-of-the-art, with respective perplexity, Hits@1 and F1 metrics of 16.28 (45 % absolute improvement), 80.7 (46 % absolute improvement) and 19.5 (20 % absolute improvement).

6 pages, 2 figures, 2 tables, NeurIPS 2018 CAI Workshop

References in corpus (3)

• A Simple, Fast Diverse Decoding Algorithm for Neural Generation
• A Deep Reinforcement Learning Chatbot
• A Hierarchical Recurrent Encoder-Decoder For Generative Context-Aware Query Suggestion

Cited by in corpus (19)

• Towards a Human-like Open-Domain Chatbot
• A Survey of Natural Language Generation Techniques with a Focus on Dialogue Systems - Past, Present and Future Directions
• Attention over Parameters for Dialogue Systems
• Neural Generation Meets Real People: Towards Emotionally Engaging Mixed-Initiative Conversations
• Bridging Text and Video: A Universal Multimodal Transformer for Video-Audio Scene-Aware Dialog
• A Pre-training Based Personalized Dialogue Generation Model with Persona-sparse Data
• A Knowledge-Enhanced Pretraining Model for Commonsense Story Generation
• End-to-End Trainable Non-Collaborative Dialog System
• MinTL: Minimalist Transfer Learning for Task-Oriented Dialogue Systems
• MEGATRON-CNTRL: Controllable Story Generation with External Knowledge Using Large-Scale Language Models
• Emotional Neural Language Generation Grounded in Situational Contexts
• CoTK: An Open-Source Toolkit for Fast Development and Fair Evaluation of Text Generation
• Generating Dialogue Responses from a Semantic Latent Space
• Semantic Role Labeling Guided Multi-turn Dialogue ReWriter
• The design and implementation of Language Learning Chatbot with XAI using Ontology and Transfer Learning
• StyleDGPT: Stylized Response Generation with Pre-trained Language Models
• Learning to Generalize for Sequential Decision Making
• Classification as Decoder: Trading Flexibility for Control in Medical Dialogue
• Improvement of a dedicated model for open domain persona-aware dialogue generation