Predicting Turn-Taking and Backchannel in Human-Machine Conversations Using Linguistic, Acoustic, and Visual Signals

TL;DR

This paper introduces a new multi-modal dataset and an end-to-end model for predicting turn-taking and backchannel actions in human-machine conversations, leveraging linguistic, acoustic, and visual signals.

Contribution

It presents an automatic data collection pipeline, a large annotated dataset, and a flexible multi-modal prediction framework that improves state-of-the-art performance.

Findings

Achieved 10% higher F1-score in turn-taking prediction.

Achieved 33% higher F1-score in backchannel prediction.

Demonstrated the effectiveness of multi-modal signals in conversation modeling.

Abstract

This paper addresses the gap in predicting turn-taking and backchannel actions in human-machine conversations using multi-modal signals (linguistic, acoustic, and visual). To overcome the limitation of existing datasets, we propose an automatic data collection pipeline that allows us to collect and annotate over 210 hours of human conversation videos. From this, we construct a Multi-Modal Face-to-Face (MM-F2F) human conversation dataset, including over 1.5M words and corresponding turn-taking and backchannel annotations from approximately 20M frames. Additionally, we present an end-to-end framework that predicts the probability of turn-taking and backchannel actions from multi-modal signals. The proposed model emphasizes the interrelation between modalities and supports any combination of text, audio, and video inputs, making it adaptable to a variety of realistic scenarios. Our experiments show that our approach achieves state-of-the-art performance on turn-taking and backchannel prediction tasks, achieving a 10% increase in F1-score on turn-taking and a 33% increase on backchannel prediction. Our dataset and code are publicly available online to ease of subsequent research.