Publications
Learning Decision Policies with Instrumental Variables through Double Machine Learning
International Conference on Machine Learning (ICML), 2024
A Universal Class of Sharpness-Aware Minimization Algorithms
International Conference on Machine Learning (ICML), 2024
High-dimensional Learning Dynamics: The Emergence of Structure and Reasoning (HiLD @ ICML workshop), 2024 (Best Paper Award)
High-dimensional Learning Dynamics: The Emergence of Structure and Reasoning (HiLD @ ICML workshop), 2024 (Best Paper Award)
On Scale-Invariant Sharpness Measures
Mathematics of Modern Machine Learning (M3L @ NeurIPS workshop), 2023
DRCFS: Doubly Robust Causal Feature Selection
International Conference on Machine Learning (ICML), 2023
Causal Feature Selection via Orthogonal Search
Transactions on Machine Learning Research (TMLR), 2022
Pyfectious: A probabilistic hierarchical simulator of infectious diseases and fine-grained control measures
PLOS Computational Biology 19(1): e1010799, 2023
GeneDisco: A Benchmark for Experimental Design in Drug Discovery
International Conference on Learning Representations (ICLR), 2022
Adaptive Experimental Design and Active Learning in the Real World (ReALML @ ICML workshop), 2022 (Spotlight Presentation)
Adaptive Experimental Design and Active Learning in the Real World (ReALML @ ICML workshop), 2022 (Spotlight Presentation)
Preprints and submissions
Faster Rates for No-Regret Learning in General Games via Cautious Optimism
submitted to The Annual ACM Symposium on Theory of Computing (STOC), 2025
Learning with Exact Invariances in Polynomial Time
submitted to International Conference on Learning Representations (ICLR), 2025
A Robust Kernel Statistical Test of Invariance: Detecting Subtle Asymmetries
submitted to International Conference on Artificial Intelligence and Statistics (AISTAT), 2025
Doubly Robust Structure Identification from Temporal Data
submitted to International Conference on Artificial Intelligence and Statistics (AISTAT), 2025
Physical Derivatives: Computing policy gradients by physical forward-propagation
submitted to International Conference on Autonomous Agents and Multi-Agent Systems (AAMAS), 2025
* denotes equal collaborations.