Graph2Speak: Improving Speaker Identification using Network Knowledge in Criminal Conversational Data

TL;DR

This paper presents Graph2Speak, a method that enhances speaker identification in criminal conversational data by leveraging social network knowledge, leading to improved accuracy in identifying speakers and conversations.

Contribution

Introducing a network-aware re-ranking approach that improves speaker identification and conversation accuracy in criminal data analysis.

Findings

Re-ranked speaker candidates based on interaction frequency.

Achieved 1.2% absolute improvement in speaker identification.

Improved conversation accuracy by 2.6% on CSI data.

Abstract

Criminal investigations mostly rely on the collection of speech conversational data in order to identify speakers and build or enrich an existing criminal network. Social network analysis tools are then applied to identify the most central characters and the different communities within the network. We introduce two candidate datasets for criminal conversational data, Crime Scene Investigation (CSI), a television show, and the ROXANNE simulated data. We also introduce the metric of conversation accuracy in the context of criminal investigations. By re-ranking candidate speakers based on the frequency of previous interactions, we improve the speaker identification baseline by 1.2% absolute (1.3% relative), and the conversation accuracy by 2.6% absolute (3.4% relative) on CSI data, and by 1.1% absolute (1.2% relative), and 2% absolute (2.5% relative) respectively on the ROXANNE simulated data.