Matches in SemOpenAlex for { <https://semopenalex.org/work/W4367000380> ?p ?o ?g. }
Showing items 1 to 63 of
63
with 100 items per page.
- W4367000380 abstract "This paper investigates policy resilience to training-environment poisoning attacks on reinforcement learning (RL) policies, with the goal of recovering the deployment performance of a poisoned RL policy. Due to the fact that the policy resilience is an add-on concern to RL algorithms, it should be resource-efficient, time-conserving, and widely applicable without compromising the performance of RL algorithms. This paper proposes such a policy-resilience mechanism based on an idea of knowledge sharing. We summarize the policy resilience as three stages: preparation, diagnosis, recovery. Specifically, we design the mechanism as a federated architecture coupled with a meta-learning manner, pursuing an efficient extraction and sharing of the environment knowledge. With the shared knowledge, a poisoned agent can quickly identify the deployment condition and accordingly recover its policy performance. We empirically evaluate the resilience mechanism for both model-based and model-free RL algorithms, showing its effectiveness and efficiency in restoring the deployment performance of a poisoned policy." @default.
- W4367000380 created "2023-04-27" @default.
- W4367000380 creator A5026148404 @default.
- W4367000380 creator A5048340011 @default.
- W4367000380 creator A5066651599 @default.
- W4367000380 date "2023-04-24" @default.
- W4367000380 modified "2023-10-16" @default.
- W4367000380 title "Policy Resilience to Environment Poisoning Attacks on Reinforcement Learning" @default.
- W4367000380 doi "https://doi.org/10.48550/arxiv.2304.12151" @default.
- W4367000380 hasPublicationYear "2023" @default.
- W4367000380 type Work @default.
- W4367000380 citedByCount "0" @default.
- W4367000380 crossrefType "posted-content" @default.
- W4367000380 hasAuthorship W4367000380A5026148404 @default.
- W4367000380 hasAuthorship W4367000380A5048340011 @default.
- W4367000380 hasAuthorship W4367000380A5066651599 @default.
- W4367000380 hasBestOaLocation W43670003801 @default.
- W4367000380 hasConcept C105339364 @default.
- W4367000380 hasConcept C111472728 @default.
- W4367000380 hasConcept C112930515 @default.
- W4367000380 hasConcept C115903868 @default.
- W4367000380 hasConcept C121332964 @default.
- W4367000380 hasConcept C138885662 @default.
- W4367000380 hasConcept C144133560 @default.
- W4367000380 hasConcept C154945302 @default.
- W4367000380 hasConcept C2779585090 @default.
- W4367000380 hasConcept C38652104 @default.
- W4367000380 hasConcept C41008148 @default.
- W4367000380 hasConcept C56739046 @default.
- W4367000380 hasConcept C89611455 @default.
- W4367000380 hasConcept C97355855 @default.
- W4367000380 hasConcept C97541855 @default.
- W4367000380 hasConceptScore W4367000380C105339364 @default.
- W4367000380 hasConceptScore W4367000380C111472728 @default.
- W4367000380 hasConceptScore W4367000380C112930515 @default.
- W4367000380 hasConceptScore W4367000380C115903868 @default.
- W4367000380 hasConceptScore W4367000380C121332964 @default.
- W4367000380 hasConceptScore W4367000380C138885662 @default.
- W4367000380 hasConceptScore W4367000380C144133560 @default.
- W4367000380 hasConceptScore W4367000380C154945302 @default.
- W4367000380 hasConceptScore W4367000380C2779585090 @default.
- W4367000380 hasConceptScore W4367000380C38652104 @default.
- W4367000380 hasConceptScore W4367000380C41008148 @default.
- W4367000380 hasConceptScore W4367000380C56739046 @default.
- W4367000380 hasConceptScore W4367000380C89611455 @default.
- W4367000380 hasConceptScore W4367000380C97355855 @default.
- W4367000380 hasConceptScore W4367000380C97541855 @default.
- W4367000380 hasLocation W43670003801 @default.
- W4367000380 hasOpenAccess W4367000380 @default.
- W4367000380 hasPrimaryLocation W43670003801 @default.
- W4367000380 hasRelatedWork W2923653485 @default.
- W4367000380 hasRelatedWork W2937603438 @default.
- W4367000380 hasRelatedWork W2952472710 @default.
- W4367000380 hasRelatedWork W2957776456 @default.
- W4367000380 hasRelatedWork W3005560120 @default.
- W4367000380 hasRelatedWork W4206669594 @default.
- W4367000380 hasRelatedWork W4221165949 @default.
- W4367000380 hasRelatedWork W4255994452 @default.
- W4367000380 hasRelatedWork W4319773215 @default.
- W4367000380 hasRelatedWork W4361026739 @default.
- W4367000380 isParatext "false" @default.
- W4367000380 isRetracted "false" @default.
- W4367000380 workType "article" @default.