ADIMA: Abuse Detection In Multilingual Audio
This addresses the lack of audio datasets for abuse detection in Indic languages, enabling content moderation in this domain, though it is incremental as it focuses on dataset creation rather than a new method.
The paper tackles the problem of detecting abusive content in multilingual audio by introducing ADIMA, a novel dataset of 11,775 audio samples in 10 Indic languages, spanning 65 hours and involving 6,446 unique users, which enables initial experiments in monolingual and cross-lingual zero-shot settings.
Abusive content detection in spoken text can be addressed by performing Automatic Speech Recognition (ASR) and leveraging advancements in natural language processing. However, ASR models introduce latency and often perform sub-optimally for profane words as they are underrepresented in training corpora and not spoken clearly or completely. Exploration of this problem entirely in the audio domain has largely been limited by the lack of audio datasets. Building on these challenges, we propose ADIMA, a novel, linguistically diverse, ethically sourced, expert annotated and well-balanced multilingual profanity detection audio dataset comprising of 11,775 audio samples in 10 Indic languages spanning 65 hours and spoken by 6,446 unique users. Through quantitative experiments across monolingual and cross-lingual zero-shot settings, we take the first step in democratizing audio based content moderation in Indic languages and set forth our dataset to pave future work.