Karan Singh

Assistant Professor of Operations Research

Bio

Karan Singh is an Assistant Professor of Operations Research in the Tepper School of Business at Carnegie Mellon University. His research centers on theory and algorithms for machine learning and optimization. His work addresses statistical and computational challenges in feedback-driven learning, spanning both prediction and control. Karan earned his PhD in Computer Science from Princeton University, where he was awarded the Porter Ogden Jacobus Fellowship, Princeton University's highest graduate student honor. Subsequently, he spent a year at Microsoft Research in Redmond as a postdoctoral researcher.

Education

• Princeton University - Ph D (Computer Science) - 2022
• Princeton University - MA (Computer Science) - 2017
• IIT Kanpur - BTech (Computer Science) - 2015

Research

Machine Learning, Control, Reinforcement Learning, Bandits, Online Learning, Convex and Nonconvex Optimization, Learning Theory

Publications

• A Boosting Approach to Reinforcement Learning

  (author(s): Nataly Bukhim, Elad Hazan, Karan Singh)
  Neural Information Processing Systems (NeurIPS), 2022

• Boosting for Online Convex Optimization

  (author(s): Elad Hazan, Karan Singh)
  International Conference on Machine Learning (ICML), 2021

• A Regret Minimization Approach to Iterative Learning Control

  (author(s): Naman Agarwal, Elad Hazan, Anirudha Majumdar, Karan Singh)
  International Conference on Machine Learning (ICML), 2021

• Improper Learning for Nonstochastic Control

  (author(s): Max Simchowitz, Karan Singh, Elad Hazan)
  Conference on Learning Theory (COLT), 2020

• No-Regret Prediction in Marginally Stable Systems

  (author(s): Udaya Ghai, Holden Lee, Karan Singh, Cyril Zhang, Yi Zhang)
  Conference on Learning Theory (COLT), 2020

• The Nonstochastic Control Problem

  (author(s): Elad Hazan, Sham Kakade, Karan Singh)
  Algorithmic Learning Theory (ALT), 2020

• Logarithmic Regret for Online Control

  (author(s): Naman Agarwal, Elad Hazan, Karan Singh)
  Neural Information Processing Systems (NeurIPS), 2019

• Online Control with Adversarial Disturbances

  (author(s): Naman Agarwal, Brian Bullins, Elad Hazan, Sham Kakade, Karan Singh)
  International Conference on Machine Learning (ICML), 2019

• Provably Efficient Maximum Entropy Exploration

  (author(s): Elad Hazan, Sham Kakade, Karan Singh, Abby Van Soest)
  International Conference on Machine Learning (ICML), 2019

• Efficient Full-Matrix Adaptive Regularization

  (author(s): Naman Agarwal, Brian Bullins, Xinyi Chen, Elad Hazan, Karan Singh, Cyril Zhang, Yi Zhang)
  International Conference on Machine Learning (ICML), 2019

• Spectral Filtering for General Linear Dynamical Systems

  (author(s): Elad Hazan, Holden Lee, Karan Singh, Cyril Zhang, Yi Zhang)
  Neural Information Processing Systems (NeurIPS), 2018

• Learning Linear Dynamical Systems via Spectral Filtering

  (author(s): Elad Hazan, Karan Singh, Cyril Zhang)
  Neural Information Processing Systems (NeurIPS), 2017

• The Price of Differential Privacy for Online Learning

  (author(s): Naman Agarwal, Karan Singh)
  International Conference on Machine Learning (ICML), 2017

• Efficient Regret Minimization in Non-Convex Games

  (author(s): Elad Hazan, Karan Singh, Cyril Zhang)
  International Conference on Machine Learning (ICML), 2017

Awards and Honors

• Princeton University, Jacobus Fellowship, highest graduate student honor (2019)
• NeurIPS, Optimization for RL workshop, Best Paper Award (2019)
• NeurIPS, Oral Presentation – top 0.5% of the submissions (2018, 2019)
• New York Academy of Sciences, 12th Annual ML Symposium, Spotlight Prize (2018)
• IIT Kanpur, President’s Gold Medal for the best academic performance (2015)

Professional Activities

• Area Chair, International Conference on Artificial Intelligence and Statistics (AISTATS) (2023)
• Program Committee, Conference on Learning Theory (COLT) (2021, 2022)
• Program Committee, Algorithmic Learning Theory (ALT) (2021, 2022)