Non-Stochastic Control

Software

deluca: A library for RL with derivatives

For a comprehensive library that provides differentiable environments, control algorithms, and benchmarking tools, check out deluca.

Preliminary notebooks to get started:

• Online Non-Stochastic Control -- Open in Colab

• Adversarial Disturbances for Controller Verification -- Open in Colab

• Regret Minimization Approach to Iterative Learning Control -- Open in GitHub

• Machine Learning for Mechanical Ventilation Control -- Open in Colab

Textbook draft on online nonstochastic control

• Website

• Arxiv version

ICML Tutorial

• Website

Publications

Initial research papers on non-stochastic control

• Online Control with Adversarial Disturbances - Naman Agarwal, Brian Bullins, Elad Hazan, Sham M. Kakade, Karan Singh - ICML 2019

• Logarithmic Regret for Control - Naman Agarwal, Elad Hazan, Karan Singh - Neurips 2019

• The Non-stochastic Control Problem - Elad Hazan, Sham M. Kakade, Karan Singh - ALT 2020

• Improper Learning for Non-stochastic Control - Max Simchowitz, Karan Singh, Elad Hazan - COLT 2020

• Adaptive Regret for Control of Time-Varying Dynamic - Paula Gradu, Elad Hazan, Edgar Minasyan

• Black-box control of linear dynamical systems - Xinyi Chen, Elad Hazan - COLT 2021

• Online Robust Control with Model Misspecification - Xinyi Chen, Udaya Ghai, Elad Hazan, Alexandre Megretski

Extensions and applications

• Deluca -- A Differentiable Control Library: Environments, Methods, and Benchmarking - Paula Gradu, John Hallman, Daniel Suo, Alex Yu, Naman Agarwal, Udaya Ghai , Karan Singh, Cyril Zhang, Aniruddha Majumdar, Elad Hazan - NeurIPS 2021 DiffCVGP Workshop

• Non-stochastic control with bandit feedback - Paula Gradu, John Hallman, Elad Hazan - Neurips 2020

• Geometric Exploration for Online Control - Orestis Pravlakis, Elad Hazan - Neurips 2020

• Boosting for Dynamical Systems - Naman Agarwal, Nataly Brukhim, Elad Hazan, Zhou Lu - ICML 2020

• A Regret Minimization Approach to Iterative Learning Control - Naman Agarwal, Elad Hazan, Karan Singh, Anirudh Majumdar - ICML 2021

• Generating Adversarial Disturbances for Controller Verification - Udaya Ghai, David Snyder, Anirudh Majumdar, Elad Hazan - L4DC 2021

• A Regret Minimization Approach to Multi-Agent Control - Udaya Ghai, Udari Madhushani, Naomi Leonard, Elad Hazan

• Provable Regret Bounds for Deep Online Learning and Control - Xinyi Chen, Edgar Minasyan, Jason D. Lee, Elad Hazan