Profile banner
Apurv Shukla

Apurv Shukla

Researcher on learning and optimization for energy and platform systems

apurv-shukla
Ann Arbor, Michigan
Joined March 2026
Favicon imageapurvshukla.github.io
Favicon imagescholar.google.com
Favicon imagegithub.com

Network

811 connections

Summary

Research focus on algorithmic decision-making for energy systems and platform economies, emphasizing learning and optimization in dynamic environments. github+2
Develops and analyzes algorithms at the intersection of bandits, optimization, and privacy—producing work on preference-based bandits, high-dimensional private bandits, and private/federated bandit frameworks. arxiv+2
Has an active academic publication record with peer-reviewed and preprint contributions (NeurIPS, ACC, arXiv) spanning power systems, learning in dynamic environments, and robust/statistical methods. neurips+2
Maintains personal and researcher profiles (personal website, Google Scholar, GitHub) and collaborates with faculty and researchers across institutions (e.g., Le Xie, P.R. Kumar, Vijay Subramanian, Debabrota Basu). github+3

Work

Education

Writing

Generalized Quantal Response Equilibrium: Existence and Efficient Learning

January 1, 2025

Introduces Generalized Quantal Response Equilibrium for bounded rational agents in general-sum games, proves existence under mild conditions, and gives efficient independent learning algorithms; arXiv preprint (2025).

Favicon imagearxiv.org

Preference-based Pure Exploration

January 1, 2024

Studies pure exploration for bandits with vector-valued rewards ordered by a preference cone, derives sample-complexity lower bounds, and proposes the Preference-based Track and Stop (PreTS) algorithm to identify the most preferred policy; presented at NeurIPS 2024.

Favicon imageproceedings.neurips.cc

Differentially Private High Dimensional Bandits

January 1, 2024

Presents PrivateLASSO, a differentially private LASSO bandit algorithm for high-dimensional contextual linear bandits, and proves privacy and utility guarantees plus minimax private lower bounds.

Favicon imagearxiv.org

Sample Complexity of Chance Constrained Optimization in Dynamic Environment

January 1, 2024

Analyzes scenario-based methods for chance-constrained optimization when scenario distributions vary over time, providing sample-complexity bounds and illustrations for convex and non-convex feasible sets; ACC 2024 submission.

Favicon imagearxiv.org

Robust Streaming PCA

January 1, 2022

Work on robust streaming principal component analysis, presented at NeurIPS 2022, addressing robustness in streaming PCA algorithms.

Favicon imageproceedings.neurips.cc