Less is More: Mitigate Spurious Correlations for Open-Domain Dialogue Response Generation Models by Causal Discovery

AbstractIn this paper, we conduct the first study on spurious correlations for open-domain response generation models based on a corpus CGDialog curated by ourselves. The current models indeed suffer from spurious correlations and have a tendency to generate irrelevant and generic re...

Full description

Bibliographic Details
Main Authors: Tao Feng, Lizhen Qu, Gholamreza Haffari
Format: Article
Language:English
Published: The MIT Press 2023-05-01
Series:Transactions of the Association for Computational Linguistics
Online Access:https://direct.mit.edu/tacl/article/doi/10.1162/tacl_a_00561/116161/Less-is-More-Mitigate-Spurious-Correlations-for