Matches in SemOpenAlex for { <https://semopenalex.org/work/W4386982011> ?p ?o ?g. }
Showing items 1 to 67 of
67
with 100 items per page.
- W4386982011 abstract "Abstract Recent work has shown that deep reinforcement learning (DRL) is vulnerable to adversarial attacks, so that exploiting vulnerabilities in DRL systems through adversarial attack techniques has become a necessary prerequisite for building robust DRL systems. Compared to traditional deep learning systems, DRL systems are characterised by long sequential decisions rather than one‐step decision, so attackers must perform multi‐step attacks on them. To successfully attack a DRL system, the number of attacks must be minimised to avoid detecting by the victim agent and to ensure the effectiveness of the attack. Some selective attack methods proposed in recent researches, that is, attacking an agent at partial time steps, are not applicable to real‐time attack scenarios, although they can avoid detecting by the victim agent. A real‐time selective attack method that is applicable to environments with discrete action spaces is proposed. Firstly, the optimal attack threshold T for performing selective attacks in the environment Env is determined. Then, the observation states corresponding to when the value of the action preference function of the victim agent in multiple eposides exceeds the threshold T are added to the training set according to this threshold. Finally, a universal perturbation is generated based on this training set, and it is used to perform real‐time selective attacks on the victim agent. Comparative experiments show that our attack method can perform real‐time attacks while maintaining the attack effect and stealthiness." @default.
- W4386982011 created "2023-09-24" @default.
- W4386982011 creator A5016645765 @default.
- W4386982011 creator A5017284019 @default.
- W4386982011 creator A5080303483 @default.
- W4386982011 creator A5088226132 @default.
- W4386982011 date "2023-09-22" @default.
- W4386982011 modified "2023-10-18" @default.
- W4386982011 title "Selective real‐time adversarial perturbations against deep reinforcement learning agents" @default.
- W4386982011 cites W2746600820 @default.
- W4386982011 cites W2963857521 @default.
- W4386982011 cites W3103780890 @default.
- W4386982011 cites W3217650841 @default.
- W4386982011 doi "https://doi.org/10.1049/cps2.12065" @default.
- W4386982011 hasPublicationYear "2023" @default.
- W4386982011 type Work @default.
- W4386982011 citedByCount "0" @default.
- W4386982011 crossrefType "journal-article" @default.
- W4386982011 hasAuthorship W4386982011A5016645765 @default.
- W4386982011 hasAuthorship W4386982011A5017284019 @default.
- W4386982011 hasAuthorship W4386982011A5080303483 @default.
- W4386982011 hasAuthorship W4386982011A5088226132 @default.
- W4386982011 hasBestOaLocation W43869820111 @default.
- W4386982011 hasConcept C121332964 @default.
- W4386982011 hasConcept C14036430 @default.
- W4386982011 hasConcept C154945302 @default.
- W4386982011 hasConcept C177264268 @default.
- W4386982011 hasConcept C199360897 @default.
- W4386982011 hasConcept C2780791683 @default.
- W4386982011 hasConcept C37736160 @default.
- W4386982011 hasConcept C38652104 @default.
- W4386982011 hasConcept C41008148 @default.
- W4386982011 hasConcept C62520636 @default.
- W4386982011 hasConcept C65856478 @default.
- W4386982011 hasConcept C78458016 @default.
- W4386982011 hasConcept C86803240 @default.
- W4386982011 hasConcept C97541855 @default.
- W4386982011 hasConceptScore W4386982011C121332964 @default.
- W4386982011 hasConceptScore W4386982011C14036430 @default.
- W4386982011 hasConceptScore W4386982011C154945302 @default.
- W4386982011 hasConceptScore W4386982011C177264268 @default.
- W4386982011 hasConceptScore W4386982011C199360897 @default.
- W4386982011 hasConceptScore W4386982011C2780791683 @default.
- W4386982011 hasConceptScore W4386982011C37736160 @default.
- W4386982011 hasConceptScore W4386982011C38652104 @default.
- W4386982011 hasConceptScore W4386982011C41008148 @default.
- W4386982011 hasConceptScore W4386982011C62520636 @default.
- W4386982011 hasConceptScore W4386982011C65856478 @default.
- W4386982011 hasConceptScore W4386982011C78458016 @default.
- W4386982011 hasConceptScore W4386982011C86803240 @default.
- W4386982011 hasConceptScore W4386982011C97541855 @default.
- W4386982011 hasLocation W43869820111 @default.
- W4386982011 hasOpenAccess W4386982011 @default.
- W4386982011 hasPrimaryLocation W43869820111 @default.
- W4386982011 hasRelatedWork W2604394466 @default.
- W4386982011 hasRelatedWork W2941205169 @default.
- W4386982011 hasRelatedWork W2952603690 @default.
- W4386982011 hasRelatedWork W3013360982 @default.
- W4386982011 hasRelatedWork W3101326751 @default.
- W4386982011 hasRelatedWork W3145910327 @default.
- W4386982011 hasRelatedWork W4225395035 @default.
- W4386982011 hasRelatedWork W4287598984 @default.
- W4386982011 hasRelatedWork W4306353320 @default.
- W4386982011 hasRelatedWork W4366126803 @default.
- W4386982011 isParatext "false" @default.
- W4386982011 isRetracted "false" @default.
- W4386982011 workType "article" @default.