Accurate and interpretable user satisfaction estimation (USE) is critical for understanding, evaluating, and continuously improving conversational systems. Users express their satisfaction or dissatisfaction with diverse conversational patterns in both general-purpose (ChatGPT and Bing Copilot) and task-oriented (customer service chatbot) conversational systems. Existing approaches based on featurized ML models or text embeddings fall short in extracting generalizable patterns and are hard to interpret. In this work, we show that LLMs can extract interpretable signals of user satisfaction from their natural language utterances more effectively than embedding-based approaches. Moreover, an LLM can be tailored for USE via an iterative prompting framework using supervision from labeled examples. The resulting method, Supervised Prompting for User satisfaction Rubrics (SPUR), not only has higher accuracy but is more interpretable as it scores user satisfaction via learned rubrics with a detailed breakdown.
@article{arxiv.2403.12388,
title = {Interpretable User Satisfaction Estimation for Conversational Systems with Large Language Models},
author = {Ying-Chun Lin and Jennifer Neville and Jack W. Stokes and Longqi Yang and Tara Safavi and Mengting Wan and Scott Counts and Siddharth Suri and Reid Andersen and Xiaofeng Xu and Deepak Gupta and Sujay Kumar Jauhar and Xia Song and Georg Buscher and Saurabh Tiwary and Brent Hecht and Jaime Teevan},
journal= {arXiv preprint arXiv:2403.12388},
year = {2024}
}