A Speaker Diarization System for Studying Peer-Led Team Learning Groups

TL;DR

This paper introduces a new unsupervised speaker diarization system for analyzing peer-led team learning sessions, utilizing a novel change detection algorithm and multi-channel data to improve accuracy and enable educational insights.

Contribution

It presents a new corpus for PLTL sessions and a novel unsupervised diarization method combining change detection and multi-channel data for better accuracy.

Findings

Improved diarization error rate over baseline systems

Effective analysis of conversational dynamics in PLTL sessions

Demonstrated utility of multi-channel audio for speaker diarization

Abstract

Peer-led team learning (PLTL) is a model for teaching STEM courses where small student groups meet periodically to collaboratively discuss coursework. Automatic analysis of PLTL sessions would help education researchers to get insight into how learning outcomes are impacted by individual participation, group behavior, team dynamics, etc.. Towards this, speech and language technology can help, and speaker diarization technology will lay the foundation for analysis. In this study, a new corpus is established called CRSS-PLTL, that contains speech data from 5 PLTL teams over a semester (10 sessions per team with 5-to-8 participants in each team). In CRSS-PLTL, every participant wears a LENA device (portable audio recorder) that provides multiple audio recordings of the event. Our proposed solution is unsupervised and contains a new online speaker change detection algorithm, termed G 3 algorithm in conjunction with Hausdorff-distance based clustering to provide improved detection accuracy. Additionally, we also exploit cross channel information to refine our diarization hypothesis. The proposed system provides good improvements in diarization error rate (DER) over the baseline LIUM system. We also present higher level analysis such as the number of conversational turns taken in a session, and speaking-time duration (participation) for each speaker.