FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W2890965997> ?p ?o ?g. }

• W2890965997 abstract "Experience replay is an important technique for addressing sample-inefficiency in deep reinforcement learning (RL), but faces difficulty in learning from binary and sparse rewards due to disproportionately few successful experiences in the replay buffer. Hindsight experience replay (HER) was recently proposed to tackle this difficulty by manipulating unsuccessful transitions, but in doing so, HER introduces a significant bias in the replay buffer experiences and therefore achieves a suboptimal improvement in sample-efficiency. In this paper, we present an analysis on the source of bias in HER, and propose a simple and effective method to counter the bias, to most effectively harness the sample-efficiency provided by HER. Our method, motivated by counter-factual reasoning and called ARCHER, extends HER with a trade-off to make rewards calculated for hindsight experiences numerically greater than real rewards. We validate our algorithm on two continuous control environments from DeepMind Control Suite - Reacher and Finger, which simulate manipulation tasks with a robotic arm - in combination with various reward functions, task complexities and goal sampling strategies. Our experiments consistently demonstrate that countering bias using more aggressive hindsight rewards increases sample efficiency, thus establishing the greater benefit of ARCHER in RL applications with limited computing budget." @default.
• W2890965997 created "2018-09-27" @default.
• W2890965997 creator A5000863703 @default.
• W2890965997 creator A5091159460 @default.
• W2890965997 date "2018-09-06" @default.
• W2890965997 modified "2023-10-01" @default.
• W2890965997 title "ARCHER: Aggressive Rewards to Counter bias in Hindsight Experience Replay." @default.
• W2890965997 cites W1498436455 @default.
• W2890965997 cites W1522301498 @default.
• W2890965997 cites W1777239053 @default.
• W2890965997 cites W1991314332 @default.
• W2890965997 cites W1992102517 @default.
• W2890965997 cites W2007754183 @default.
• W2890965997 cites W2012987867 @default.
• W2890965997 cites W2100677568 @default.
• W2890965997 cites W2120179935 @default.
• W2890965997 cites W2139053308 @default.
• W2890965997 cites W2140135625 @default.
• W2890965997 cites W2141559645 @default.
• W2890965997 cites W2145339207 @default.
• W2890965997 cites W2149860990 @default.
• W2890965997 cites W2158782408 @default.
• W2890965997 cites W2165150801 @default.
• W2890965997 cites W2201581102 @default.
• W2890965997 cites W2290354866 @default.
• W2890965997 cites W2295899596 @default.
• W2890965997 cites W2328078142 @default.
• W2890965997 cites W2335959470 @default.
• W2890965997 cites W2614839826 @default.
• W2890965997 cites W2733961795 @default.
• W2890965997 cites W2772721022 @default.
• W2890965997 cites W2781585732 @default.
• W2890965997 cites W2788781499 @default.
• W2890965997 cites W2950492145 @default.
• W2890965997 cites W2964161785 @default.
• W2890965997 cites W2964299886 @default.
• W2890965997 hasPublicationYear "2018" @default.
• W2890965997 type Work @default.
• W2890965997 sameAs 2890965997 @default.
• W2890965997 citedByCount "7" @default.
• W2890965997 countsByYear W28909659972019 @default.
• W2890965997 countsByYear W28909659972020 @default.
• W2890965997 countsByYear W28909659972021 @default.
• W2890965997 crossrefType "posted-content" @default.
• W2890965997 hasAuthorship W2890965997A5000863703 @default.
• W2890965997 hasAuthorship W2890965997A5091159460 @default.
• W2890965997 hasConcept C10347200 @default.
• W2890965997 hasConcept C119857082 @default.
• W2890965997 hasConcept C154945302 @default.
• W2890965997 hasConcept C15744967 @default.
• W2890965997 hasConcept C162324750 @default.
• W2890965997 hasConcept C175444787 @default.
• W2890965997 hasConcept C180747234 @default.
• W2890965997 hasConcept C185592680 @default.
• W2890965997 hasConcept C198531522 @default.
• W2890965997 hasConcept C2778869765 @default.
• W2890965997 hasConcept C2779458634 @default.
• W2890965997 hasConcept C41008148 @default.
• W2890965997 hasConcept C43617362 @default.
• W2890965997 hasConcept C77805123 @default.
• W2890965997 hasConcept C97541855 @default.
• W2890965997 hasConceptScore W2890965997C10347200 @default.
• W2890965997 hasConceptScore W2890965997C119857082 @default.
• W2890965997 hasConceptScore W2890965997C154945302 @default.
• W2890965997 hasConceptScore W2890965997C15744967 @default.
• W2890965997 hasConceptScore W2890965997C162324750 @default.
• W2890965997 hasConceptScore W2890965997C175444787 @default.
• W2890965997 hasConceptScore W2890965997C180747234 @default.
• W2890965997 hasConceptScore W2890965997C185592680 @default.
• W2890965997 hasConceptScore W2890965997C198531522 @default.
• W2890965997 hasConceptScore W2890965997C2778869765 @default.
• W2890965997 hasConceptScore W2890965997C2779458634 @default.
• W2890965997 hasConceptScore W2890965997C41008148 @default.
• W2890965997 hasConceptScore W2890965997C43617362 @default.
• W2890965997 hasConceptScore W2890965997C77805123 @default.
• W2890965997 hasConceptScore W2890965997C97541855 @default.
• W2890965997 hasLocation W28909659971 @default.
• W2890965997 hasOpenAccess W2890965997 @default.
• W2890965997 hasPrimaryLocation W28909659971 @default.
• W2890965997 hasRelatedWork W1130790960 @default.
• W2890965997 hasRelatedWork W2121863487 @default.
• W2890965997 hasRelatedWork W2145339207 @default.
• W2890965997 hasRelatedWork W2736601468 @default.
• W2890965997 hasRelatedWork W2789008106 @default.
• W2890965997 hasRelatedWork W2795921402 @default.
• W2890965997 hasRelatedWork W2892979040 @default.
• W2890965997 hasRelatedWork W2897007337 @default.
• W2890965997 hasRelatedWork W2911276197 @default.
• W2890965997 hasRelatedWork W2920768993 @default.
• W2890965997 hasRelatedWork W2964001908 @default.
• W2890965997 hasRelatedWork W2964048876 @default.
• W2890965997 hasRelatedWork W2970776272 @default.
• W2890965997 hasRelatedWork W2996110979 @default.
• W2890965997 hasRelatedWork W3087618584 @default.
• W2890965997 hasRelatedWork W3095991084 @default.
• W2890965997 hasRelatedWork W3102583453 @default.
• W2890965997 hasRelatedWork W3112023111 @default.
• W2890965997 hasRelatedWork W3187882526 @default.
• W2890965997 hasRelatedWork W567721252 @default.
• W2890965997 isParatext "false" @default.

• next