An Empirical Investigation of Pre-Trained Transformer Language Models for Open-Domain Dialogue Generation

TL;DR

This paper empirically evaluates pre-trained Transformer language models for open-domain dialogue generation, analyzing their performance across multiple datasets, languages, and decoding strategies to improve relevance and diversity.

Contribution

It introduces a joint prediction paradigm for context and response, and provides comprehensive experimental analysis of various models and decoding methods in dialogue generation.

Findings

Transformer models outperform baselines in relevance and diversity

Joint prediction improves response quality

Decoding strategies significantly affect generated responses

Abstract

We present an empirical investigation of pre-trained Transformer-based auto-regressive language models for the task of open-domain dialogue generation. Training paradigm of pre-training and fine-tuning is employed to conduct the parameter learning. Corpora of News and Wikipedia in Chinese and English are collected for the pre-training stage respectively. Dialogue context and response are concatenated into a single sequence utilized as the input of the models during the fine-tuning stage. A weighted joint prediction paradigm for both context and response is designed to evaluate the performance of models with or without the loss term for context prediction. Various of decoding strategies such as greedy search, beam search, top-k sampling, etc. are employed to conduct the response text generation. Extensive experiments are conducted on the typical single-turn and multi-turn dialogue corpora such as Weibo, Douban, Reddit, DailyDialog, and Persona-Chat. Detailed numbers of automatic evaluation metrics on relevance and diversity of the generated results for the languages models as well as the baseline approaches are reported.