CoLLAB: A Collaborative Approach for Multilingual Abuse Detection

TL;DR

This paper introduces CoLLAB, a novel framework for multilingual audio abuse detection that merges models across languages without retraining, improving scalability and performance in diverse linguistic environments.

Contribution

The paper proposes CoLLAB, a new model merging approach that enables multilingual abuse detection without additional training, addressing scalability and resource challenges.

Findings

PTM representations outperform others in AAD

Combining PTM representations improves accuracy

CoLLAB achieves competitive multilingual AAD performance

Abstract

In this study, we investigate representations from paralingual Pre-Trained model (PTM) for Audio Abuse Detection (AAD), which has not been explored for AAD. Our results demonstrate their superiority compared to other PTM representations on the ADIMA benchmark. Furthermore, combining PTM representations enhances AAD performance. Despite these improvements, challenges with cross-lingual generalizability still remain, and certain languages require training in the same language. This demands individual models for different languages, leading to scalability, maintenance, and resource allocation issues and hindering the practical deployment of AAD systems in linguistically diverse real-world environments. To address this, we introduce CoLLAB, a novel framework that doesn't require training and allows seamless merging of models trained in different languages through weight-averaging. This results in a unified model with competitive AAD performance across multiple languages.