Matches in SemOpenAlex for { <https://semopenalex.org/work/W3046308277> ?p ?o ?g. }
Showing items 1 to 93 of
93
with 100 items per page.
- W3046308277 abstract "Performance and safety are often two competing objectives in decision-making problems. We study the problem of integrating a collection of controllers with different safety and performance levels into one that takes a middle-ground position amongst them. In the first contribution, we formulate the problem of blending controllers using the framework of constrained Markov decision processes and contextual multi-objective bandits. We use the reward function and the auxiliary costs of the Markov decision process to measure the performance and the safety of a controller, respectively. We subsequently use these measures to form the feedback of a bandit whose arms are the input controllers. The blending algorithm must interact with the bandit and minimize a regret term that measures the suboptimality of the pulled arms with respect to an expert whose choice of arms is Pareto optimal. In the second contribution, we design a blending algorithm and show that its average regret converges to zero. We also derive an upper bound on the algorithm’s suboptimality in performance and safety and we show that its computation imposes no additional computational complexity. We empirically demonstrate the algorithm’s success in blending a safe and a performant controller in a variety of Safety Gym environments. The results reflect the following key takeaway: the blended controller shows a strict improvement in performance compared to the safe controller and is safer than the performant controller." @default.
- W3046308277 created "2020-08-07" @default.
- W3046308277 creator A5042227524 @default.
- W3046308277 creator A5050718717 @default.
- W3046308277 creator A5068441112 @default.
- W3046308277 creator A5086083363 @default.
- W3046308277 date "2022-06-08" @default.
- W3046308277 modified "2023-10-16" @default.
- W3046308277 title "Blending Controllers via Multi-Objective Bandits" @default.
- W3046308277 cites W1995053973 @default.
- W3046308277 cites W2057847320 @default.
- W3046308277 cites W2060846151 @default.
- W3046308277 cites W2061753713 @default.
- W3046308277 cites W2095602422 @default.
- W3046308277 cites W2097746659 @default.
- W3046308277 cites W2113818300 @default.
- W3046308277 cites W2773218293 @default.
- W3046308277 cites W2787908307 @default.
- W3046308277 cites W2796061040 @default.
- W3046308277 cites W2909027522 @default.
- W3046308277 cites W2913181250 @default.
- W3046308277 cites W2963525569 @default.
- W3046308277 cites W2963575966 @default.
- W3046308277 cites W2963751010 @default.
- W3046308277 cites W2965754719 @default.
- W3046308277 cites W3103262232 @default.
- W3046308277 cites W3104371626 @default.
- W3046308277 cites W4206530644 @default.
- W3046308277 cites W4250589301 @default.
- W3046308277 doi "https://doi.org/10.23919/acc53348.2022.9867486" @default.
- W3046308277 hasPublicationYear "2022" @default.
- W3046308277 type Work @default.
- W3046308277 sameAs 3046308277 @default.
- W3046308277 citedByCount "1" @default.
- W3046308277 countsByYear W30463082772021 @default.
- W3046308277 crossrefType "proceedings-article" @default.
- W3046308277 hasAuthorship W3046308277A5042227524 @default.
- W3046308277 hasAuthorship W3046308277A5050718717 @default.
- W3046308277 hasAuthorship W3046308277A5068441112 @default.
- W3046308277 hasAuthorship W3046308277A5086083363 @default.
- W3046308277 hasBestOaLocation W30463082772 @default.
- W3046308277 hasConcept C105795698 @default.
- W3046308277 hasConcept C106189395 @default.
- W3046308277 hasConcept C11413529 @default.
- W3046308277 hasConcept C119857082 @default.
- W3046308277 hasConcept C126255220 @default.
- W3046308277 hasConcept C136197465 @default.
- W3046308277 hasConcept C154945302 @default.
- W3046308277 hasConcept C159886148 @default.
- W3046308277 hasConcept C203479927 @default.
- W3046308277 hasConcept C2775924081 @default.
- W3046308277 hasConcept C33923547 @default.
- W3046308277 hasConcept C41008148 @default.
- W3046308277 hasConcept C45374587 @default.
- W3046308277 hasConcept C47446073 @default.
- W3046308277 hasConcept C50817715 @default.
- W3046308277 hasConcept C6557445 @default.
- W3046308277 hasConcept C86803240 @default.
- W3046308277 hasConceptScore W3046308277C105795698 @default.
- W3046308277 hasConceptScore W3046308277C106189395 @default.
- W3046308277 hasConceptScore W3046308277C11413529 @default.
- W3046308277 hasConceptScore W3046308277C119857082 @default.
- W3046308277 hasConceptScore W3046308277C126255220 @default.
- W3046308277 hasConceptScore W3046308277C136197465 @default.
- W3046308277 hasConceptScore W3046308277C154945302 @default.
- W3046308277 hasConceptScore W3046308277C159886148 @default.
- W3046308277 hasConceptScore W3046308277C203479927 @default.
- W3046308277 hasConceptScore W3046308277C2775924081 @default.
- W3046308277 hasConceptScore W3046308277C33923547 @default.
- W3046308277 hasConceptScore W3046308277C41008148 @default.
- W3046308277 hasConceptScore W3046308277C45374587 @default.
- W3046308277 hasConceptScore W3046308277C47446073 @default.
- W3046308277 hasConceptScore W3046308277C50817715 @default.
- W3046308277 hasConceptScore W3046308277C6557445 @default.
- W3046308277 hasConceptScore W3046308277C86803240 @default.
- W3046308277 hasLocation W30463082771 @default.
- W3046308277 hasLocation W30463082772 @default.
- W3046308277 hasOpenAccess W3046308277 @default.
- W3046308277 hasPrimaryLocation W30463082771 @default.
- W3046308277 hasRelatedWork W2016425266 @default.
- W3046308277 hasRelatedWork W2122187689 @default.
- W3046308277 hasRelatedWork W2150379124 @default.
- W3046308277 hasRelatedWork W2157016390 @default.
- W3046308277 hasRelatedWork W2161367706 @default.
- W3046308277 hasRelatedWork W3110641737 @default.
- W3046308277 hasRelatedWork W3111617249 @default.
- W3046308277 hasRelatedWork W3176362036 @default.
- W3046308277 hasRelatedWork W4287555357 @default.
- W3046308277 hasRelatedWork W4315575041 @default.
- W3046308277 isParatext "false" @default.
- W3046308277 isRetracted "false" @default.
- W3046308277 magId "3046308277" @default.
- W3046308277 workType "article" @default.