Abhishek Varghese

Abhishek Varghese

PhD Student, UMass Amherst | ML for Healthcare & Wearable Sensing

Publications

You can also find my articles on my Google Scholar profile.

Journal Articles

Bio-Inspired Self-Supervised Learning for Wrist-worn IMU Signals

Published in arXiv preprint, 2026

We introduce a novel tokenization strategy grounded in the submovement theory of motor control for learning robust human-activity representations from wrist accelerometers. A Transformer encoder pretrained via masked movement-segment reconstruction on the NHANES corpus (~28k hours, ~11k participants) outperforms strong wearable SSL baselines across six subject-disjoint HAR benchmarks.

Recommended citation: Tarale, P., Chu, K., Varghese, A., Liu, K., Xu, M. A., Iyyer, M., & Lee, S. I. (2026). "Bio-Inspired Self-Supervised Learning for Wrist-worn IMU Signals." arXiv:2603.10961.
Download Paper

QEDCartographer: Automating Formal Verification Using Reward-Free Reinforcement Learning

Published in Proceedings of the IEEE/ACM International Conference on Software Engineering (ICSE), 2025

QEDCartographer is an automated proof-synthesis tool that combines supervised and reinforcement learning to explore the proof space for Coq theorems. By incorporating proofs' branching structure and enabling reward-free search, it overcomes the sparse-reward problem in formal verification and proves 21.4% of CoqGym benchmark theorems — outperforming all prior single-tool approaches.

Recommended citation: Sanchez-Stern, A., Varghese, A., Kaufman, Z., Zhang, D., Ringer, T., & Brun, Y. (2025). "QEDCartographer: Automating Formal Verification Using Reward-Free Reinforcement Learning." Proceedings of ICSE 2025. Ottawa, Ontario, Canada.
Download Paper