μBoost: An Effective Method for Solving Indic Multilingual Text Classification Problem

Text Classification is an integral part of many Natural Language Processing tasks such as sarcasm detection, sentiment analysis and many more such applications. Many e-commerce websites, social-media/entertainment platforms use such models to enhance user-experience to generate traffic and thus revenue on their platforms. In this paper, we are presenting our solution to Multilingual Abusive Comment Identification Challenge on Moj, an Indian video-sharing social networking service, powered by ShareChat and IIIT-Delhi for the 2nd Workshop on Emerging Advances in Multimodal AI (EAM) at IEEE BigMM, 2021. The challenge dealt with detecting abusive comments, in 13 regional Indic languages, on the videos on Moj platform. Our solution utilizes the novel mu Boost, an ensemble of CatBoost classifier models and Multilingual Representations for Indian Languages (MURIL) model, to produce SOTA performance on Indic text classification tasks. We were able to achieve a mean F1-score of 89.286 on the test data, an improvement over baseline MURIL model with a F1-score of 87.48.