DLRG@TamilNLP-ACL2022: Offensive Span Identification in Tamil usingBiLSTM-CRF approach

Abstract

Identifying offensive speech is an exciting and essential area of research, with ample traction in recent times. This paper presents our system submission to the subtask 1, focusing on using supervised approaches for extracting Offensive spans from code-mixed Tamil-English comments. To identify offensive spans, we developed the Bidirectional Long Short-Term Memory (BiLSTM) model with Glove Embedding. With this method, the developed system achieved an overall F1 of 0.1728. Additionally, for comments with less than 30 characters, the developed system shows an F1 of 0.3890, competitive with other submissions.