Log In Sign Up

DialoGPT: Large-Scale Generative Pre-training for Conversational Response Generation

by   Yizhe Zhang, et al.

We present a large, tunable neural conversational response generation model, DialoGPT (dialogue generative pre-trained transformer). Trained on 147M conversation-like exchanges extracted from Reddit comment chains over a period spanning from 2005 through 2017, DialoGPT extends the Hugging Face PyTorch transformer to attain a performance close to human both in terms of automatic and human evaluation in single-turn dialogue settings. We show that conversational systems that leverage DialoGPT generate more relevant, contentful and context-consistent responses than strong baseline systems. The pre-trained model and training pipeline are publicly released to facilitate research into neural response generation and the development of more intelligent open-domain dialogue systems.


EVA: An Open-Domain Chinese Dialogue System with Large-Scale Generative Pre-Training

Although pre-trained language models have remarkably enhanced the genera...

Vector Representations of Idioms in Conversational Systems

We demonstrate, in this study, that an open-domain conversational system...

Context Matters in Semantically Controlled Language Generation for Task-oriented Dialogue Systems

This work combines information about the dialogue history encoded by pre...

Detecting Interlocutor Confusion in Situated Human-Avatar Dialogue: A Pilot Study

In order to enhance levels of engagement with conversational systems, ou...

Learning Locality and Isotropy in Dialogue Modeling

Existing dialogue modeling methods have achieved promising performance o...

ConceptNet infused DialoGPT for Underlying Commonsense Understanding and Reasoning in Dialogue Response Generation

The pre-trained conversational models still fail to capture the implicit...

On Task-Level Dialogue Composition of Generative Transformer Model

Task-oriented dialogue systems help users accomplish tasks such as booki...