Publications Peer-reviewed papers, preprints, and work under review. Google Scholar. Under review OracleZoom: On-Policy Self-Distillation Inspired Reference-Constrained Recursive Image Super Resolution Under review Shubhashis Roy Dipta†, Sourajit Saha†, Shaswati Saha, Nobin Sarwar† Equal contribution arXivCodeModelsSiteDemo DecomposeRL: Learning to Ask Useful, Informative, and Diverse Questions for Semi-Supervised, Traceable Claim Verification Under review Shubhashis Roy Dipta, Ankur Padia, Francis Ferraro arXivCodeModelsSite Omni-Modal Dissonance Benchmark: Systematically Breaking Modality Consensus to Probe Robustness and Calibrated Abstention Under review Zabir Al Nazi†, Shubhashis Roy Dipta†, Md Rizwan Parvez† Equal contribution arXiv Last Translation Benchmark Under review Vilém Zouhar, Niyati Bafna, Mukund Choudhary, …, Shubhashis Roy Dipta, … and 240 more arXivSite The Backdrop Exposes What the World Around an Agent Costs It Under review Nusrat Jahan Lia, Shubhashis Roy Dipta arXivCodeSite Did the Problem Actually Get Harder? Testing LLMs on Paired P and NP-Complete Problems Hardened by Adversarial Multi-Agent Rewriting Under review Nurul Labib Sayeedi, Md. Faiyaz Abdullah Sayeedi, Zabir Al Nazi, Mahbub E Sobhani, Shubhashis Roy Dipta arXivCode CulturalRiddles: A Multicultural Benchmark of Riddles Under review My Chiffon Nguyen, Tianyi Hu, Vinod Anbalagan, …, Shubhashis Roy Dipta, … and 160 more arXivCodeDatasetSite ChannelGuard: From Borrowed Safety to Application-Controlled Defenses in Multi-Agent LLMs Under review Elias Hossain, Md Mehedi Hasan Nipu, Fatema Tuj Johora Faria, Shubhashis Roy Dipta, Maleeha Sheikh, Tasfia Nuzhat Ornee, Niloofar Yousefi arXiv 2026 Cornerstones or Stumbling Blocks? Deciphering the Rock Tokens in On-Policy Distillation NeurIPS 2026 Yuxuan Jiang†, Runchao Li†, Shubhashis Roy Dipta†, Dawei Li, Zhao Yang† Equal contribution arXivCode TRIAGE: Evaluating Prospective Metacognitive Control in LLMs under Resource Constraints NeurIPS 2026 Zabir Al Nazi, Shubhashis Roy Dipta arXiv AgentCollabBench: Diagnosing When Good Agents Make Bad Collaborators NeurIPS 2026 Aritra Mazumder, Shubhashis Roy Dipta, Nusrat Jahan Lia, Tanzila Khan, Kainat Raisa Hossain, Nehaa Shri, Shubhrangshu Debsarkar, Humayra Tasnim, Gour Gupal Talukder Shawon, Debjoty Mitra, Sumaiya Ahmed Rani, Al Jami Islam Anik, Al Nafeu Khan arXivCodeSite PA3: Policy-Aware Agent Alignment through Chain-of-Thought EMNLP 2026 Shubhashis Roy Dipta, Daniel Bis, Kun Zhou, Lichao Wang, Benjamin Z. Yao, Chenlei Guo, Ruhi SarikayaWork done during internship at Amazon Alexa AI arXivModelsSlidesVideo †DAGGER: Distractor-Aware Graph Generation for Executable Reasoning in Math Problems EMNLP 2026 Zabir Al Nazi, Shubhashis Roy Dipta, Sudipta Kar arXivCodeSite Many Dialects, Many Languages, One Cultural Lens: Evaluating Multilingual VLMs for Bengali Culture Understanding Across Historically Linked Languages and Regional Dialects EMNLP 2026 Nurul Labib Sayeedi, Md. Faiyaz Abdullah Sayeedi, Shubhashis Roy Dipta, Mahbub E Sobhani, Rubaya Tabassum, Ariful Ekraj Hridoy, Mehraj Mahmood, Md. Tarek Hasan, Swakkhar Shatabda arXivCodeDatasetSite Register Shifts Break LLM Safety: A Bengali Benchmark with Culturally Grounded Harms EMNLP 2026 Naymul Islam, Nusrat Jahan Lia, Shubhashis Roy Dipta, Sabik Bin Sultan, Abdullah Khan Zehady arXivCodeDatasetSite GanitLLM: Difficulty-Aware Bengali Mathematical Reasoning through Curriculum-GRPO ACL 2026 Shubhashis Roy Dipta, Khairul Mahbub, Nadia Najjar arXivCodeModelsSitePosterVideo VC-Inspector: Advancing Reference-free Evaluation of Video Captions with Factual Analysis ACL 2026 Shubhashis Roy Dipta, Tz-Ying Wu, Subarna Tripathi arXivCodeModelsSiteSlidesPosterVideo Multimodal Unlearning Across Vision, Language, Video, and Audio: Survey of Methods, Datasets, and Benchmarks ACL 2026 Nobin Sarwar, Shubhashis Roy Dipta, Zheyuan Liu, Vaidehi Patil arXivPDFCodeSite 2025 Q2E: Query-to-Event Decomposition for Zero-Shot Multilingual Text-to-Video Retrieval AACL 2025 Shubhashis Roy Dipta, Francis Ferraro arXivCodeModelsSiteSlidesPosterVideo If We May De-Presuppose: Robustly Verifying Claims through Presupposition-Free Question Decomposition *SEM 2025 Shubhashis Roy Dipta, Francis Ferraro arXivCodeSitePoster Learning How to Use Tools, Not Just When: Pattern-Aware Tool-Integrated Reasoning MathAI @NeurIPS 2025 Ningning Xu, Yuxuan Jiang, Shubhashis Roy Dipta arXiv 2024 MethEvo: an accurate evolutionary information-based methylation site predictor Springer 2024 Sadia Islam, Shafayat Bin Shabbir Mugdha, Shubhashis Roy Dipta, Md. Easin Arafat, Swakkhar Shatabda, Hamid Alinejad-Rokny, Iman Dehzangi HU at SemEval-2024 Task 8A: Can Contrastive Learning Learn Embeddings to Detect Machine-Generated Text? SemEval 2024 Shubhashis Roy Dipta, Sadat Shahriar Code UMBCLU at SemEval-2024 Task 1: Semantic Textual Relatedness with and without machine translation SemEval 2024 Shubhashis Roy Dipta, Sai Vallurupalli Code 2023 Semantically-informed Hierarchical Event Modeling *SEM 2023 Shubhashis Roy Dipta, Mehdi Rezaee, Francis Ferraro arXivCodeSlides 2020 Mal-Light: Enhancing Lysine Malonylation Sites Prediction Problem Using Evolutionary-based Features IEEE Access 2020 Md. Wakil Ahmad, Md. Easin Arafat, Ghazaleh Taherzadeh, Alok Sharma, Shubhashis Roy Dipta, Abdollah Dehzangi, Swakkhar Shatabda SEMal: Accurate protein malonylation site predictor using structural and evolutionary information Elsevier 2020 Shubhashis Roy Dipta, Ghazaleh Taherzadeh, Md. Wakil Ahmad, Md. Easin Arafat, Swakkhar Shatabda, Abdollah Dehzangi Code