Multilingual Abusive Comment Detection is an innovative new challenge towards combating abusive comments on Moj, one of India's largest short-video apps in multiple regional languages. The target label is 1 or 0 denoting whether a particular comment is abusive or not respectively. In this project we have used the following network architecture:
- The Embedding layer of size 32
- The LSTM Layer with 100 memory inputs
We obtained an accuracy of 76.67% on the test set. The architecture is pretty primitive and we need to handle different languages into consideration. The work is in progress.