Research.

2026

• A Machine Learning and Explainable AI-Based Multiclass Police Fraud Prediction Scheme with SHAP Based Interpretability

  Publication / ICECTE 2026, IEEE

  M. Chowdhury, J. Islam, Md. A. I. Semon, S. Sarker, M. M. Barua, A. Akter

  Peer reviewed. SHAP attribution on every prediction, so a flagged case traces back to the features that drove it.

• Still Loading@DravidianLangTech 2026: Telugu Prompt-Style Recovery using Multilingual Transformers

  Publication / Proceedings of DravidianLangTech, ACL 2026 / Code

  Samonwita Sarker, Isnat Mehrin Sami, Priyontee Mojumder, Arpita Mallik, Hasan Murad

  First author. Sorting Telugu transcripts into nine communicative styles, comparing four multilingual transformers under focal loss. The MuRIL system placed 7th at 0.1703 macro F1, where chance sits near 0.11.