Rating Prediction in Conversational Task Assistants with Behavioral and Conversational-Flow Features

TL;DR

This paper introduces TB-Rater, a Transformer-based model that combines conversational-flow and user behavior features to predict user ratings in conversational task assistants, demonstrating improved accuracy on real Alexa TaskBot data.

Contribution

The paper presents a novel Transformer model that integrates behavioral and conversational-flow features for rating prediction in CTA, validated on real multimodal conversational data.

Findings

Model outperforms baselines in rating prediction accuracy.

Combining behavioral and conversational features yields better insights.

Behavioral features can bootstrap future CTA systems.

Abstract

Predicting the success of Conversational Task Assistants (CTA) can be critical to understand user behavior and act accordingly. In this paper, we propose TB-Rater, a Transformer model which combines conversational-flow features with user behavior features for predicting user ratings in a CTA scenario. In particular, we use real human-agent conversations and ratings collected in the Alexa TaskBot challenge, a novel multimodal and multi-turn conversational context. Our results show the advantages of modeling both the conversational-flow and behavioral aspects of the conversation in a single model for offline rating prediction. Additionally, an analysis of the CTA-specific behavioral features brings insights into this setting and can be used to bootstrap future systems.