Mamta
King's College London
I am a postdoctoral researcher in the Department of Informatics at King’s College London, where I work with Dr. Oana Cocarascu on enhancing the robustness of NLP systems. My work is dedicated to making AI more inclusive and accessible. I earned my Ph.D. from the AI-NLP-ML Lab at the Indian Institute of Technology Patna, with a thesis titled “Advancing Sentiment Analysis: From Diverse Domains to Multilinguality and Robustness.”
My research interests cover several key areas: Robustness of uni-modal and multi-modal AI systems, Fact Verification, Multilinguality, Multilingual reasoning in Indian languages, Model Interpretability, Fairness in Machine Learning, and opinion mining.
news
| Oct 22, 2026 | Co-organizing a shared task on multilingual AI safety to be held in conjunction with ICON 2026 (December 2026, Guwahati, India). More details available here. |
|---|---|
| Sep 01, 2026 | 🎉 Two papers accepted at EMNLP 2026: IndicSafeEval: “Safety Robustness of Large Language Models under Multilingual Persuasive Jailbreak Attacks” (Findings) and “To Trust or Not to Trust: Retrieval-Augmented Fact Checking in Speech” (Main). |
| Jul 27, 2026 | Co-organizing the Workshop on Foundation Models for Social Good at IJCAI-ECAI 2026, which will be held on 16 August 2026 in Bremen, Germany. More details about the workshop are available here. |
| May 21, 2026 | 🎉 Our paper MLingualFC: Evaluating Jailbreak Vulnerabilities in Multilingual Vision-Language Models has been accepted at MeLLM, ACL 2026. |
| Apr 07, 2026 | 🎉 Our paper TinyAttack: Exploring Stylistic Vulnerabilities in Large Language Models has been accepted at ACL 2026 Findings! |
| Mar 01, 2026 | 📄 Two papers accepted at VarDial 2026 (co-located with EACL 2026): Improving Dialect Robustness in LLMs and Indic-TunedLens: Interpreting Multilingual Models in Indian Languages. |
|---|---|
| Nov 01, 2025 | 🏆 Our paper I-GUARD: Interpretability-Guided Parameter Optimization for Adversarial Defense has been accepted at EMNLP 2025 Findings! |
| Apr 01, 2025 | 📄 Our paper Federated Model Synchronization for Diagnostic Redefinition through a Novel Selective Parameter Unlearning has been accepted at WACV 2026! |
| Jan 15, 2024 | 📄 Our paper FactEval: Evaluating the Robustness of Fact Verification Systems in the Era of Large Language Models has been accepted at NAACL 2025! |