Sep 2026 Six papers accepted at AACL-IJCNLP 2026 — The Language of Bargaining (main), Navigating the Reality Gap: Privacy-Preserving On-Device Continual Adaptation of ASR for Clinical Telephony (main), Anchored Turns (main), Method Ranking Inverts Across Targets (main), When Reject Turns into Accept (findings), and Benchmarking Frontier Agentic Deception in LLMs vs Humans (findings). View ↗
Sep 2026 Two papers accepted at EMNLP 2026 Workshop REALM — LLM Graders Are Systematically Harsher Than Industry Interviewers, and CricBench: A Multilingual Benchmark for Evaluating LLMs in Cricket Analytics. View ↗
Sep 2026 Beyond Accuracy accepted at EMNLP 2026 Workshop on Mathematical NLP (MathNLP) — nine-dimension framework for diagnosing algebraic reasoning failures in LLMs. Paper ↗
Aug 2026 Six papers accepted at EMNLP 2026 — Format Collapse in Policy Optimization for LLM Reasoning (main), "I Strongly Suspect This Website Is a Scam" (main), OrgAccess (main), FinBalance: A Multi-Document Accounting Reconciliation Benchmark (findings), PRAXIS (findings), and Certified Cost-Optimal Triage for Suggestion Detection in Reviews (industry track). View ↗
Jul 2026 Evaluating and Mitigating Misgendering in English-to-Hindi Machine Translation accepted at COLM 2026. View ↗
Jul 2026 Oracle-Agnostic Text Quality Optimization via Prompt-Only Agentic Loops accepted at CAISc 2026 — sole-authored by Dr. Dhruv Kumar. Details ↗
Jun 2026 Measuring Representation Robustness in Large Language Models for Geometry accepted at ICML 2026 Workshop on AI for Math (AI4Math). View ↗
May 2026 Five Manan Gupta–authored papers accepted across ICML 2026 workshops — DPMI: A Principled Index for Neural Polysemanticity, Diagnosing LLM Judge Reliability, Is the Trust Region Necessary?, Latent Phase-Shift Rollback, and Reasoning Phases Are Continuous, Not Discrete. View ↗
May 2026 Six time-series foundation model papers accepted across ICML 2026 workshops (FMSD, GlobalSouthML, Forecast) — TSQueryBench, GITCO, Mix, Don't Pick, ReGeN, Semantics or Structure?, and Where Computation Lives Inside TabPFN. View ↗
May 2026 MultiVulnBench and four more reasoning benchmarks — Evaluating LLM Reasoning on Operating System Algorithms, GraphStateEval, Joint Evaluation of Compliance, Planning, and Consistency, and The Complexity Ceiling Benchmark — accepted at ICML 2026 Workshop CTB. View ↗
May 2026 MafiaPersona and PoolBench accepted at ICML 2026 Workshop NExT-Game. View ↗
May 2026 Bridging the Visual Gap (Meitei Mayek), Evaluating Indic Literary Creativity, and IndicDB accepted at ICML 2026 Workshop GlobalSouthML. View ↗
May 2026 Authority, Truth, and Citation Bias, What Do Lorentz-Equivariant Jet Taggers Learn?, and Context Over Content accepted across ICML 2026 workshops (EIML, AI4GOOD, AI4Physics, AgenticUQ, SPIGM, FoGen, FAGEN). View ↗
Apr 2026 Beyond Accuracy preprint — nine-dimension framework for diagnosing algebraic reasoning failures in LLMs. Paper ↗
Apr 2026 LUDOBENCH preprint — benchmark for evaluating LLM strategic reasoning through Ludo board game scenarios. Paper ↗
Apr 2026 LLM-as-a-Judge for Time Series Explanations preprint — evaluating factual correctness of LLM-generated time series explanations. Paper ↗
Apr 2026 BLEND accepted at The Web Conference (WWW 2026) — balanced and leaf-enhanced dual fine-tuning for taxonomy completion. View ↗
Apr 2026 From Trust to Compromise accepted at ACL 2026 (Main Conference) — outcome-verified LLM phishing simulation and real-time defense. View ↗
Apr 2026 Two papers accepted at ICLR 2026 Workshop on LLM Reasoning — The Language of Bargaining: Linguistic Effects in LLM Negotiations, and From Facts to Conclusions: Integrating Deductive Reasoning in Retrieval-Augmented LLMs. View ↗
Mar 2026 PhantomGrad receives ₹34.8L grant from Gujarat Police for SPANDAN — an AI/ML project to strengthen police investigation processes and improve conviction outcomes, under an MoU between Gujarat Police and BITS Pilani. View ↗
Mar 2026 Roamify accepted at The Web Conference (WWW 2026) — LLM-based personalised itinerary planning Chrome extension. Paper ↗
Mar 2026 RASE: Retain-Agnostic Machine Unlearning via Activation Subspace Projection accepted at CVPR 2026 Workshop on Machine Unlearning for Visual Understanding (MUV). View ↗
Jan 2026 Sakshm AI accepted at ACM Journal on Computing and Sustainable Societies (JCSS 2026) — Socratic AI tutor for coding education in India. Paper ↗
Jan 2026 A Hybrid Supervised-LLM Pipeline for Actionable Suggestion Mining in Unstructured Customer Reviews accepted at EACL 2026 Industry Track. Paper ↗
Nov 2025 Loop Invariant Generation paper wins Best Paper Award at AIWare 2025. Paper ↗
Oct 2025 Two papers accepted at EMNLP 2025 — Pedagogical Teacher-Student LLM Agents (main conference) and ReviewEval (findings). View ↗
Sep 2025 time2time paper accepted at NeurIPS 2025 Workshop on Recent Advances in Time Series Foundation Models (BERT2S). Paper ↗
Sep 2025 Talking with Oompa Loompas accepted at NeurIPS 2025 Workshops on WiML and LLM Evaluation. Paper ↗
Aug 2025 Rubric Is All You Need accepted at ACM ICER 2025 — improving LLM-based code evaluation with question-specific rubrics. Paper ↗
May 2025 The Impact of LLMs on K-12 Education in Rural India published in Scientific Reports. Paper ↗
Mar 2025 The Role of Generative AI in Mechanical Engineering Education published in Scientific Reports. Paper ↗
Feb 2025 DebateBench accepted at NAACL Student Research Workshop 2025. Paper ↗
Feb 2025 Analyzing LLM Usage in an Advanced Computing Class accepted at ACE 2025. Paper ↗
Nov 2024 ₹20L grant awarded by BITS Pilani for "Revolutionizing Indian Higher Education Institutes with Generative AI" (Nov 2024 – Oct 2026). Details ↗
Aug 2024 ₹30L grant from WISE–IIE Higher Education Consortium for Generative AI in Education research. Details ↗
Aug 2024 ₹25L grant from The Nature Conservancy for AI for River–Climate Modeling. Details ↗
May 2024 Dr. Dhruv Kumar joins BITS Pilani as Assistant Professor and founds PhantomGrad. People ↗
Jan 2024 Google Cloud Education Grant of ₹12.5L awarded to support AI/LLM research at BITS Pilani. Details ↗
May 2023 AggFirstJoin wins Best Paper Award at IEEE/ACM CCGrid 2023. Paper ↗