TransferTransfo: A Transfer Learning Approach for Neural Network Based Conversational Agents
arXiv:1901.08149
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 (9)
- A Neural Conversational Model
- How NOT To Evaluate Your Dialogue System: An Empirical Study of Unsupervised Evaluation Metrics for Dialogue Response Generation
- A Simple, Fast Diverse Decoding Algorithm for Neural Generation
- Adversarial Learning for Neural Dialogue Generation
- A Deep Reinforcement Learning Chatbot
- Personalizing Dialogue Agents: I have a dog, do you have pets too?
- Set Transformer: A Framework for Attention-based Permutation-Invariant Neural Networks
- Neural Net Models for Open-Domain Discourse Coherence
- A Hierarchical Recurrent Encoder-Decoder For Generative Context-Aware Query Suggestion
Cited by in corpus (105)
- Towards a Human-like Open-Domain Chatbot
- Poly-encoders: Transformer Architectures and Pre-training Strategies for Fast and Accurate Multi-sentence Scoring
- SpinalNet: Deep Neural Network with Gradual Input
- DialoGPT: Large-Scale Generative Pre-training for Conversational Response Generation
- SOLOIST: Building Task Bots at Scale with Transfer Learning and Machine Teaching
- ACUTE-EVAL: Improved Dialogue Evaluation with Optimized Questions and Multi-turn Comparisons
- One Chatbot Per Person: Creating Personalized Chatbots based on Implicit User Profiles
- Towards an Online Empathetic Chatbot with Emotion Causes
- Challenges in Building Intelligent Open-domain Dialog Systems
- EmpTransfo: A Multi-head Transformer Architecture for Creating Empathetic Dialog Systems
- CharacterChat: Supporting the Creation of Fictional Characters through Conversation and Progressive Manifestation with a Chatbot
- Open-Domain Conversational Agents: Current Progress, Open Problems, and Future Directions
- A Survey of Natural Language Generation Techniques with a Focus on Dialogue Systems - Past, Present and Future Directions
- Task-Oriented Dialogue System as Natural Language Generation
- Alternating Recurrent Dialog Model with Large-scale Pre-trained Language Models
- GenNI: Human-AI Collaboration for Data-Backed Text Generation
- Exploring Prompt-based Few-shot Learning for Grounded Dialog Generation
- Recent Advances and Challenges in Task-oriented Dialog System
- Neural Generation Meets Real People: Towards Emotionally Engaging Mixed-Initiative Conversations
- Attention over Parameters for Dialogue Systems
- CAiRE: An Empathetic Neural Chatbot
- Neural Language Generation: Formulation, Methods, and Evaluation
- Bridging Text and Video: A Universal Multimodal Transformer for Video-Audio Scene-Aware Dialog
- Unstructured Text Enhanced Open-domain Dialogue System: A Systematic Survey
- You Impress Me: Dialogue Generation via Mutual Persona Perception
- A Pre-training Based Personalized Dialogue Generation Model with Persona-sparse Data
- Profile Consistency Identification for Open-domain Dialogue Agents
- A Knowledge-Enhanced Pretraining Model for Commonsense Story Generation
- Getting To Know You: User Attribute Extraction from Dialogues
- PLATO: Pre-trained Dialogue Generation Model with Discrete Latent Variable
- Pretrained Language Models for Text Generation: A Survey
- BoB: BERT Over BERT for Training Persona-based Dialogue Models from Limited Personalized Data
- MinTL: Minimalist Transfer Learning for Task-Oriented Dialogue Systems
- End-to-End Trainable Non-Collaborative Dialog System
- MEGATRON-CNTRL: Controllable Story Generation with External Knowledge Using Large-Scale Language Models
- Policy-Driven Neural Response Generation for Knowledge-Grounded Dialogue Systems
- Retrieval-Augmented Transformer-XL for Close-Domain Dialog Generation
- Towards Persona-Based Empathetic Conversational Models
- Are Pre-trained Language Models Knowledgeable to Ground Open Domain Dialogues?
- A Tailored Pre-Training Model for Task-Oriented Dialog Generation
- CheerBots: Chatbots toward Empathy and Emotionusing Reinforcement Learning
- Knowledge Injection into Dialogue Generation via Language Models
- Emotional Neural Language Generation Grounded in Situational Contexts
- MoEL: Mixture of Empathetic Listeners
- TopicRefine: Joint Topic Prediction and Dialogue Response Generation for Multi-turn End-to-End Dialogue System
- Action-Based Conversations Dataset: A Corpus for Building More In-Depth Task-Oriented Dialogue Systems
- Controllable Abstractive Dialogue Summarization with Sketch Supervision
- Audrey: A Personalized Open-Domain Conversational Bot
- Set-to-Sequence Methods in Machine Learning: a Review
- CoTK: An Open-Source Toolkit for Fast Development and Fair Evaluation of Text Generation
- Proto: A Neural Cocktail for Generating Appealing Conversations
- A Stack-Propagation Framework for Low-Resource Personalized Dialogue Generation
- Knowledge-Grounded Dialogue Generation with Pre-trained Language Models
- A Large-Scale Chinese Short-Text Conversation Dataset
- Designing Precise and Robust Dialogue Response Evaluators
- Hybrid Generative-Retrieval Transformers for Dialogue Domain Adaptation
- Sparse Text Generation
- Will I Sound Like Me? Improving Persona Consistency in Dialogues through Pragmatic Self-Consciousness
- Exploring Versatile Generative Language Model Via Parameter-Efficient Transfer Learning
- Generating Personalized Dialogue via Multi-Task Meta-Learning
- Learning to Retrieve Entity-Aware Knowledge and Generate Responses with Copy Mechanism for Task-Oriented Dialogue Systems
- Partner Matters! An Empirical Study on Fusing Personas for Personalized Response Selection in Retrieval-Based Chatbots
- Rome was built in 1776: A Case Study on Factual Correctness in Knowledge-Grounded Response Generation
- Weakly-Supervised Open-Retrieval Conversational Question Answering
- Distilling BERT for low complexity network training
- Video-Grounded Dialogues with Pretrained Generation Language Models
- Argument Undermining: Counter-Argument Generation by Attacking Weak Premises
- Open-Domain Dialogue Generation Based on Pre-trained Language Models
- Generating Dialogue Responses from a Semantic Latent Space
- On Learning Text Style Transfer with Direct Rewards
- Dialogue-oriented Pre-training
- Measuring Conversational Uptake: A Case Study on Student-Teacher Interactions
- Empathetic Dialogue Generation with Pre-trained RoBERTa-GPT2 and External Knowledge
- Towards a Universal NLG for Dialogue Systems and Simulators with Future Bridging
- CloneBot: Personalized Dialogue-Response Predictions
- Learning to Predict Persona Information forDialogue Personalization without Explicit Persona Description
- Commonsense-Focused Dialogues for Response Generation: An Empirical Study
- The design and implementation of Language Learning Chatbot with XAI using Ontology and Transfer Learning
- Phrase-BERT: Improved Phrase Embeddings from BERT with an Application to Corpus Exploration
- Local Knowledge Powered Conversational Agents
- Pretrained Language Models for Dialogue Generation with Multiple Input Sources
- A Simple and Efficient Multi-Task Learning Approach for Conditioned Dialogue Generation
- Semantic Role Labeling Guided Multi-turn Dialogue ReWriter
- EmpBot: A T5-based Empathetic Chatbot focusing on Sentiments
- Probing Task-Oriented Dialogue Representation from Language Models
- StyleDGPT: Stylized Response Generation with Pre-trained Language Models
- Proposal Towards a Personalized Knowledge-powered Self-play Based Ensemble Dialog System
- Do You Have the Right Scissors? Tailoring Pre-trained Language Models via Monte-Carlo Methods
- Conversational Semantic Role Labeling
- A Corpus of Controlled Opinionated and Knowledgeable Movie Discussions for Training Neural Conversation Models
- Cross-lingual Intermediate Fine-tuning improves Dialogue State Tracking
- Learning to Generalize for Sequential Decision Making
- Transferable Persona-Grounded Dialogues via Grounded Minimal Edits
- Data-Efficient Methods for Dialogue Systems
- Injecting Hierarchy with U-Net Transformers
- Classification as Decoder: Trading Flexibility for Control in Medical Dialogue
- AffectON: Incorporating Affect Into Dialog Generation
- Improvement of a dedicated model for open domain persona-aware dialogue generation
- Local Explanation of Dialogue Response Generation
- Multi-Granularity Representations of Dialog
- Grounding Conversations with Improvised Dialogues
- End-to-End Learning of Flowchart Grounded Task-Oriented Dialogs
- Partner Personas Generation for Diverse Dialogue Generation
- Estimating Subjective Crowd-Evaluations as an Additional Objective to Improve Natural Language Generation
- Findings from Experiments of On-line Joint Reinforcement Learning of Semantic Parser and Dialogue Manager with real Users