Matches in SemOpenAlex for { <https://semopenalex.org/work/W3132858839> ?p ?o ?g. }
- W3132858839 abstract "We study black-box reward poisoning attacks against reinforcement learning (RL), in which an adversary aims to manipulate the rewards to mislead a sequence of RL agents with unknown algorithms to learn a nefarious policy in an environment unknown to the adversary a priori. That is, our attack makes minimum assumptions on the prior knowledge of the adversary: it has no initial knowledge of the environment or the learner, and neither does it observe the learner's internal mechanism except for its performed actions. We design a novel black-box attack, U2, that can provably achieve a near-matching performance to the state-of-the-art white-box attack, demonstrating the feasibility of reward poisoning even in the most challenging black-box setting." @default.
- W3132858839 created "2021-03-01" @default.
- W3132858839 creator A5022094334 @default.
- W3132858839 creator A5027711113 @default.
- W3132858839 creator A5068386240 @default.
- W3132858839 creator A5073837056 @default.
- W3132858839 date "2021-02-16" @default.
- W3132858839 modified "2023-09-24" @default.
- W3132858839 title "Reward Poisoning in Reinforcement Learning: Attacks Against Unknown Learners in Unknown Environments." @default.
- W3132858839 cites W1633675443 @default.
- W3132858839 cites W1771410628 @default.
- W3132858839 cites W1777239053 @default.
- W3132858839 cites W1850488217 @default.
- W3132858839 cites W1931078822 @default.
- W3132858839 cites W1988526405 @default.
- W3132858839 cites W2020764470 @default.
- W3132858839 cites W2103104707 @default.
- W3132858839 cites W2115293355 @default.
- W3132858839 cites W2132876566 @default.
- W3132858839 cites W2132908009 @default.
- W3132858839 cites W2159600763 @default.
- W3132858839 cites W2168565265 @default.
- W3132858839 cites W2182055801 @default.
- W3132858839 cites W21934178 @default.
- W3132858839 cites W2410842990 @default.
- W3132858839 cites W2410983263 @default.
- W3132858839 cites W2518564545 @default.
- W3132858839 cites W2604394466 @default.
- W3132858839 cites W2605788860 @default.
- W3132858839 cites W2616841723 @default.
- W3132858839 cites W2786676179 @default.
- W3132858839 cites W2794908222 @default.
- W3132858839 cites W2890679110 @default.
- W3132858839 cites W2890752237 @default.
- W3132858839 cites W2902572901 @default.
- W3132858839 cites W2944955921 @default.
- W3132858839 cites W2946619602 @default.
- W3132858839 cites W2949103145 @default.
- W3132858839 cites W2949608212 @default.
- W3132858839 cites W2955689724 @default.
- W3132858839 cites W2962755762 @default.
- W3132858839 cites W2962820691 @default.
- W3132858839 cites W2962948945 @default.
- W3132858839 cites W2963049774 @default.
- W3132858839 cites W2963068985 @default.
- W3132858839 cites W2963207607 @default.
- W3132858839 cites W2963289505 @default.
- W3132858839 cites W2963308241 @default.
- W3132858839 cites W2963582321 @default.
- W3132858839 cites W2964054583 @default.
- W3132858839 cites W2964138440 @default.
- W3132858839 cites W2964299116 @default.
- W3132858839 cites W2966120739 @default.
- W3132858839 cites W2970259165 @default.
- W3132858839 cites W2970502043 @default.
- W3132858839 cites W2970734210 @default.
- W3132858839 cites W2970912396 @default.
- W3132858839 cites W3004977066 @default.
- W3132858839 cites W3013223143 @default.
- W3132858839 cites W3034593529 @default.
- W3132858839 cites W3034871777 @default.
- W3132858839 cites W3035244686 @default.
- W3132858839 cites W3035388736 @default.
- W3132858839 cites W3035515732 @default.
- W3132858839 cites W3082960758 @default.
- W3132858839 cites W3099159407 @default.
- W3132858839 cites W3177136857 @default.
- W3132858839 cites W950880443 @default.
- W3132858839 hasPublicationYear "2021" @default.
- W3132858839 type Work @default.
- W3132858839 sameAs 3132858839 @default.
- W3132858839 citedByCount "2" @default.
- W3132858839 countsByYear W31328588392020 @default.
- W3132858839 countsByYear W31328588392021 @default.
- W3132858839 crossrefType "posted-content" @default.
- W3132858839 hasAuthorship W3132858839A5022094334 @default.
- W3132858839 hasAuthorship W3132858839A5027711113 @default.
- W3132858839 hasAuthorship W3132858839A5068386240 @default.
- W3132858839 hasAuthorship W3132858839A5073837056 @default.
- W3132858839 hasConcept C105795698 @default.
- W3132858839 hasConcept C111472728 @default.
- W3132858839 hasConcept C11413529 @default.
- W3132858839 hasConcept C138885662 @default.
- W3132858839 hasConcept C154945302 @default.
- W3132858839 hasConcept C15744967 @default.
- W3132858839 hasConcept C165064840 @default.
- W3132858839 hasConcept C33923547 @default.
- W3132858839 hasConcept C38652104 @default.
- W3132858839 hasConcept C41008148 @default.
- W3132858839 hasConcept C41065033 @default.
- W3132858839 hasConcept C48103436 @default.
- W3132858839 hasConcept C67203356 @default.
- W3132858839 hasConcept C75553542 @default.
- W3132858839 hasConcept C77805123 @default.
- W3132858839 hasConcept C94966114 @default.
- W3132858839 hasConcept C97541855 @default.
- W3132858839 hasConceptScore W3132858839C105795698 @default.
- W3132858839 hasConceptScore W3132858839C111472728 @default.
- W3132858839 hasConceptScore W3132858839C11413529 @default.
- W3132858839 hasConceptScore W3132858839C138885662 @default.