Towards Cross-Lingual Audio Abuse Detection in Low-Resource Settings with Few-Shot Learning

Aditya Narayan Sankaran, Reza Farahbakhsh, Noel Crespi
12/2/2024
cs.CLcs.AI

No official code found. Best implementation: callmesanfornow/fsl-audio-abuse (1 star, MIT, last commit 2025-01-21).

Abstract

Online abusive content detection, particularly in low-resource settings and within the audio modality, remains underexplored. We investigate the potential of pre-trained audio representations for detecting abusive language in low-resource languages, in this case, in Indian languages using Few Shot Learning (FSL). Leveraging powerful representations from models such as Wav2Vec and Whisper, we explore cross-lingual abuse detection using the ADIMA dataset with FSL. Our approach integrates these representations within the Model-Agnostic Meta-Learning (MAML) framework to classify abusive language in 10 languages. We experiment with various shot sizes (50-200) evaluating the impact of limited data on performance. Additionally, a feature visualization study was conducted to better understand model behaviour. This study highlights the generalization ability of pre-trained models in low-resource scenarios and offers valuable insights into detecting abusive language in multilingual contexts.

View on arXivView PDF

Code Implementations(1)

callmesanfornow/fsl-audio-abuseCommunity100%
MIT

Code repository for COLING 2025 Paper on Few Shot Audio Abuse Detection in Low Resource Languages

10Python, Jupyter NotebookDec 2, 20241 years agoMIT

Similar Papers

Deep Learning Models for Multilingual Hate Speech Detection

85% match1 repo1122020

Bridging Modalities: Enhancing Cross-Modality Hate Speech Detection with Few-Shot In-Context Learning

84% match2024

Creating and Evaluating Code-Mixed Nepali-English and Telugu-English Datasets for Abusive Language Detection Using Traditional and Deep Learning Models

84% match2025

Can Prompting LLMs Unlock Hate Speech Detection across Languages? A Zero-shot and Few-shot Study

84% match2025

Vicinal Risk Minimization for Few-Shot Cross-lingual Transfer in Abusive Language Detection

84% match2023

A Multi-Task Benchmark for Abusive Language Detection in Low-Resource Settings

83% match1 repo102025

Cite this paper

@article{sankaran2024towards,
  title  = {Towards Cross-Lingual Audio Abuse Detection in Low-Resource Settings with Few-Shot Learning},
  author = {Aditya Narayan Sankaran and Reza Farahbakhsh and Noel Crespi},
  year   = {2024},
  eprint = {2412.01408},
  archivePrefix = {arXiv},
  url    = {http://arxiv.org/abs/2412.01408v3}
}

Add a badge to your README

Shows whether official code is linked for this paper.

Code of Paper badge
[![Code of Paper](https://codeofpaper.com/badge/2412.01408.svg)](https://codeofpaper.com/paper/2412.01408)

Discussion