On the Reliability of User Feedback for Evaluating the Quality of Conversational Agents
Jordan Massiah, Emine Yılmaz, Yunlong Jiao, Gabriella Kazai · 2023
We analyse the reliability of users' explicit feedback for evaluating the quality of conversational agents. Using data from a commercial conversational system, we analyse how user feedback compares with human annotations; how well it aligns with implicit user satisfaction signals, such as retention; and how much user feedback is needed to reliably evaluate the quality of a conversational system.