English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Safely exploring an unknown dynamical system is critical to the deployment ofreinforcement learning (RL) in physical systems where failures may havecatastrophic consequences. In scenarios where one knows little about thedynamics, diverse transition data covering relevant regions of state-actionspace is needed to apply either model-based or model-free RL. Motivated by thecooling of Google's data centers, we study how one can safely identify theparameters of a system model with a desired accuracy and confidence level. Inparticular, we focus on learning an unknown linear system with Gaussian noiseassuming only that, initially, a nominal safe action is known. Define safety assatisfying specific linear constraints on the state space (e.g., requirementson process variable) that must hold over the span of an entire trajectory, andgiven a Probably Approximately Correct (PAC) style bound on the estimationerror of model parameters, we show how to compute safe regions of action spaceby gradually growing a ball around the nominal safe action. One can apply anyexploration strategy where actions are chosen from such safe regions.Experiments on a stylized model of data center cooling dynamics show howcomputing proper safe regions can increase the sample efficiency of safeexploration.

Safe Exploration for Identifying Linear Systems via Robust Optimization