DG
Debargha Ganguly
cs.LGcs.AIcs.CLcs.CVcs.LOcs.DCcs.CRcs.ITcs.NEcs.RO
On Valency
published · living versionsW_mmefdpxh·v1 · currentpublished
LEANGUARD: Lean-Verified Guardrails for Tool-Using Agents
with Sreehari Sankar, Vikash Singh, Mohsen Hariri, Weicong Chen +15
1 version
W_u8955kt7·v1published
W_etmeccs6·v1published
Preprints & journals
23 papers in the corpus · 2024–2026Sequent-Prover: Training Agents for Formal, Checkable SMT-based Reasoning2610.15258v1 · Debargha Ganguly, Weicong Chen, Mohsen Hariri et al.2026 · 0 citationsarXiv
Correcting WHERE, Preserving HOW: Compositional Generalization for Vision-Language-Action Models via Referential Guidance2609.38616v1 · Yanyan Zhang, Disheng Liu, Xinpeng Li et al.2026 · 0 citationsarXiv
How to Loop MoE: Flatten the Experts, Untie the Attention2609.35751v1 · Shouren Wang, Chuang Ma, Mohsen Hariri et al.2026 · 0 citationsarXiv
Beyond Solver Verdicts: Generative Reward Models for Autoformalization2609.11085v2 · Vikash Singh, Debargha Ganguly, Aman Goel et al.2026 · 0 citationsarXiv
StochBench: A Domain-Specific Benchmark for Stochastic Processes in Lean2609.09264v1 · Idan Davidovich, Debargha Ganguly, Vikash Singh et al.2026 · 0 citationsarXiv
Auditing Cross-Lingual Fairness in Language Model Watermarking2608.20047v1 · Alexander Nemecek, Osama Zafar, Debargha Ganguly et al.2026 · 0 citationsarXiv
Test-Time Scaling in Reasoning LLMs: Inference Regimes, Evaluation, and Reproducibility2608.04001v2 · Mohsen Hariri, Weicong Chen, Nahal Shahini et al.2026 · 0 citationsarXiv
CausalGuard: Conformal Inference under Graph Uncertainty2605.21928v1 · Vikash Singh, Weicong Chen, Debargha Ganguly et al.2026 · 0 citationsarXiv
Privacy Policy Enforcement Guardrails for Data-Sensitive Retrieval-Augmented Generation2605.17034v2 · Osama Zafar, Alexander Nemecek, Yiqian Zhang et al.2026 · 0 citationsarXiv
Reliability-Gated Source Anchoring for Continual Test-Time Adaptation2605.14063v2 · Vikash Singh, Debargha Ganguly, Weicong Chen et al.2026 · 0 citationsarXiv
Path-Lock Expert: Separating Reasoning Mode in Hybrid Thinking via Architecture-Level Separation2604.27201v3 · Shouren Wang, Wang Yang, Chuang Ma et al.2026 · 0 citationsarXiv
AgentCE-Bench: Agent Configurable Evaluation with Scalable Horizons and Controllable Difficulty under Lightweight Environments2604.06111v2 · Wang Yang, Chaoda Song, Xinpeng Li et al.2026 · 0 citationsarXiv
Toward Guarantees for Clinical Reasoning in Vision Language Models via Formal Verification2602.24111v2 · Vikash Singh, Debargha Ganguly, Haotian Yu et al.2026 · 0 citationsarXiv
Context Determines Optimal Architecture in Materials Segmentation2602.04154v1 · Mingjian Lu, Pawan K. Tripathi, Mark Shteyn et al.2026 · 0 citationsarXiv
Trust The Typical2602.04581v1 · Debargha Ganguly, Sreehari Sankar, Biyao Zhang et al.2026 · 0 citationsarXiv
Mid-Think: Training-Free Intermediate-Budget Reasoning via Token-Level Triggers2601.07036v2 · Wang Yang, Debargha Ganguly, Xinpeng Li et al.2026 · 0 citationsarXiv
Demystifying Hybrid Thinking: Can LLMs Truly Switch Between Think and No-Think?2510.12680v2 · Shouren Wang, Wang Yang, Xianxuan Long et al.2025 · 0 citationsarXiv
Efficient Fine-Grained GPU Performance Modeling for Distributed Deep Learning of LLM2509.22832v1 · Biyao Zhang, Mingkai Zheng, Debargha Ganguly et al.2025 · 0 citationsarXiv
: Online Log Anomaly Detection Via Unsupervised Typicality Learning2507.20051v1 · Weicong Chen, Vikash Singh, Zahra Rahmani et al.2025 · 2 citations2025 IEEE 32nd International Conference on High Performance Computing, Data, and Analytics (HiPC), pp. 96-107, 2025
Grammars of Formal Uncertainty: When to Trust LLMs in Automated Reasoning Tasks2505.20047v1 · Debargha Ganguly, Vikash Singh, Sreehari Sankar et al.2025 · 2 citationsarXiv
Forte : Finding Outliers with Representation Typicality Estimation2410.01322v2 · Debargha Ganguly, Warren Morningstar, Andrew Yu et al.2024 · 0 citationsarXiv
Visual Concept Networks: A Graph-Based Approach to Detecting Anomalous Data in Deep Neural Networks2409.18235v1 · Debargha Ganguly, Debayan Gupta, Vipin Chaudhary2024 · 0 citationsarXiv
Proof of Thought : Neurosymbolic Program Synthesis allows Robust and Interpretable Reasoning2409.17270v2 · Debargha Ganguly, Srinivasan Iyengar, Vipin Chaudhary et al.2024 · 0 citationsarXivon Valency
Career total: 29 works. 23 are in this corpus.
Profile built from the corpus for this byline.
Author records are still filling in while the Hub is in alpha. If this is your page, you'll be able to claim it soon. Spot a mistake? Tell us.