Logarithmic Regret for Adversarial Online Control

Authors: Dylan Foster, Max Simchowitz

ICML 2020 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Theoretical Our main result is to achieve logarithmic regret for fully adversarial disturbances, provided that costs are known and quadratic. Our algorithm and analysis use a characterization for the optimal offline control law to reduce the online control problem to (delayed) online learning with approximate advantage functions.
Researcher Affiliation Academia 1Massachusetts Institute of Technology 2UC Berkeley. Correspondence to: Dylan Foster <dylanf@mit.edu>.
Pseudocode Yes Algorithm 1 Riccatitron; Algorithm 2 Online Newton Step (ONS(ε,η,C,Σ)); Algorithm 3
Open Source Code No The paper does not provide any concrete access (link, explicit statement of release for their code) to open-source code for the methodology described.
Open Datasets No The paper is theoretical and does not mention or use any datasets.
Dataset Splits No The paper is theoretical and does not mention any training, validation, or test dataset splits.
Hardware Specification No The paper does not provide specific hardware details used for running experiments.
Software Dependencies No The paper does not provide specific ancillary software details with version numbers.
Experiment Setup No The paper does not contain specific experimental setup details, hyperparameters, or training configurations.