Evaluating Open-Domain Dialogues in Latent Space with Next Sentence Prediction and Mutual Information

Kun Zhao, Bohao Yang, Chenghua Lin, Wenge Rong, Aline Villavicencio +1 more
5/26/2023
cs.CL

No official code found. Best implementation: Bernard-Yang/CMN-ACL2023 (4 stars, last commit 2024-03-14).

Abstract

The long-standing one-to-many issue of the open-domain dialogues poses significant challenges for automatic evaluation methods, i.e., there may be multiple suitable responses which differ in semantics for a given conversational context. To tackle this challenge, we propose a novel learning-based automatic evaluation metric (CMN), which can robustly evaluate open-domain dialogues by augmenting Conditional Variational Autoencoders (CVAEs) with a Next Sentence Prediction (NSP) objective and employing Mutual Information (MI) to model the semantic similarity of text in the latent space. Experimental results on two open-domain dialogue datasets demonstrate the superiority of our method compared with a wide range of baselines, especially in handling responses which are distant to the golden reference responses in semantics.

View on arXivView PDF

Code Implementations(1)

Bernard-Yang/CMN-ACL2023Community100%
42PythonMay 5, 20232 years ago

Similar Papers

SLIDE: A Framework Integrating Small and Large Language Models for Open-Domain Dialogues Evaluation

87% match1 repo32024

DiQAD: A Benchmark Dataset for End-to-End Open-domain Dialogue Assessment

84% match2023

RADE: Reference-Assisted Dialogue Evaluation for Open-Domain Dialogue

84% match1 repo52023

Emphasising Structured Information: Integrating Abstract Meaning Representation into LLMs for Enhanced Open-Domain Dialogue Evaluation

84% match1 repo12024

PairEval: Open-domain Dialogue Evaluation with Pairwise Comparison

84% match1 repo52024

Simple LLM Prompting is State-of-the-Art for Robust and Multilingual Dialogue Evaluation

84% match1 repo42023

Cite this paper

@article{zhao2023evaluating,
  title  = {Evaluating Open-Domain Dialogues in Latent Space with Next Sentence Prediction and Mutual Information},
  author = {Kun Zhao and Bohao Yang and Chenghua Lin and Wenge Rong and Aline Villavicencio and Xiaohui Cui},
  year   = {2023},
  eprint = {2305.16967},
  archivePrefix = {arXiv},
  url    = {http://arxiv.org/abs/2305.16967v3}
}

Add a badge to your README

Shows whether official code is linked for this paper.

Code of Paper badge
[![Code of Paper](https://codeofpaper.com/badge/2305.16967.svg)](https://codeofpaper.com/paper/2305.16967)

Discussion