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 Tools annual

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 Tools monthly

Global: Zendy Plus monthly

Presents the access of premium content as premium feature

Premium Content

Global: Zendy Plus annual

Global: Zendy Free Plan

Deep reinforcement learning algorithms (DRL) are increasingly being used insafety-critical systems. Ensuring the safety of DRL agents is a criticalconcern in such contexts. However, relying solely on testing is not sufficientto ensure safety as it does not offer guarantees. Building safety monitors isone solution to alleviate this challenge. This paper proposes SMARLA, a machinelearning-based safety monitoring approach designed for DRL agents. Forpractical reasons, SMARLA is designed to be black-box (as it does not requireaccess to the internals of the agent) and leverages state abstraction to reducethe state space and thus facilitate the learning of safety violation predictionmodels from agent's states. We validated SMARLA on two well-known RL casestudies. Empirical analysis reveals that SMARLA achieves accurate violationprediction with a low false positive rate, and can predict safety violations atan early stage, approximately halfway through the agent's execution beforeviolations occur.

SMARLA: A Safety Monitoring Approach for Deep Reinforcement Learning Agents