SyntaxMind at BLP-2025 Task 1: Leveraging Attention Fusion of CNN and GRU for Hate Speech Detection

TL;DR

This paper presents a novel hate speech detection system for Bangla text that combines CNN and GRU architectures with attention mechanisms, achieving high accuracy in the BLP-2025 challenge.

Contribution

It introduces a unified model integrating BanglaBERT embeddings with CNN and GRU branches, enhancing contextual and local feature capture for hate speech classification.

Findings

Achieved 0.7345 micro F1-Score in Subtask 1A

Achieved 0.7317 micro F1-Score in Subtask 1B

Secured 2nd and 5th places in respective subtasks

Abstract

This paper describes our system used in the BLP-2025 Task 1: Hate Speech Detection. We participated in Subtask 1A and Subtask 1B, addressing hate speech classification in Bangla text. Our approach employs a unified architecture that integrates BanglaBERT embeddings with multiple parallel processing branches based on GRUs and CNNs, followed by attention and dense layers for final classification. The model is designed to capture both contextual semantics and local linguistic cues, enabling robust performance across subtasks. The proposed system demonstrated high competitiveness, obtaining 0.7345 micro F1-Score (2nd place) in Subtask 1A and 0.7317 micro F1-Score (5th place) in Subtask 1B.