Rethinking the Evaluation of Dialogue Systems: Effects of User Feedback on Crowdworkers and LLMs

TL;DR

This paper investigates how user feedback, explicit or implicit, affects the evaluation of task-oriented dialogue systems by crowdworkers and LLMs, revealing significant differences and implications for future assessment methods.

Contribution

It introduces a comparative methodology for evaluating dialogue systems with and without user feedback, highlighting its impact on assessment consistency and personalization.

Findings

User feedback influences evaluation ratings significantly.

Crowdworkers are more affected by user feedback on usefulness and interestingness.

User feedback improves agreement among crowdworkers on complex requests.

Abstract

In ad-hoc retrieval, evaluation relies heavily on user actions, including implicit feedback. In a conversational setting such signals are usually unavailable due to the nature of the interactions, and, instead, the evaluation often relies on crowdsourced evaluation labels. The role of user feedback in annotators' assessment of turns in a conversational perception has been little studied. We focus on how the evaluation of task-oriented dialogue systems (TDSs), is affected by considering user feedback, explicit or implicit, as provided through the follow-up utterance of a turn being evaluated. We explore and compare two methodologies for assessing TDSs: one includes the user's follow-up utterance and one without. We use both crowdworkers and large language models (LLMs) as annotators to assess system responses across four aspects: relevance, usefulness, interestingness, and explanation quality. Our findings indicate that there is a distinct difference in ratings assigned by both annotator groups in the two setups, indicating user feedback does influence system evaluation. Workers are more susceptible to user feedback on usefulness and interestingness compared to LLMs on interestingness and relevance. User feedback leads to a more personalized assessment of usefulness by workers, aligning closely with the user's explicit feedback. Additionally, in cases of ambiguous or complex user requests, user feedback improves agreement among crowdworkers. These findings emphasize the significance of user feedback in refining system evaluations and suggest the potential for automated feedback integration in future research. We publicly release the annotated data to foster research in this area.