Matches in SemOpenAlex for { <https://semopenalex.org/work/W3010717779> ?p ?o ?g. }
- W3010717779 abstract "To operate effectively in the real world, agents should be able to act from high-dimensional raw sensory input such as images and achieve diverse goals across long time-horizons. Current deep reinforcement and imitation learning methods can learn directly from high-dimensional inputs but do not scale well to long-horizon tasks. In contrast, classical graphical methods like A* search are able to solve long-horizon tasks, but assume that the state space is abstracted away from raw sensory input. Recent works have attempted to combine the strengths of deep learning and classical planning; however, dominant methods in this domain are still quite brittle and scale poorly with the size of the environment. We introduce Sparse Graphical Memory (SGM), a new data structure that stores states and feasible transitions in a sparse memory. SGM aggregates states according to a novel two-way consistency objective, adapting classic state aggregation criteria to goal-conditioned RL: two states are redundant when they are interchangeable both as goals and as starting states. Theoretically, we prove that merging nodes according to two-way consistency leads to an increase in shortest path lengths that scales only linearly with the merging threshold. Experimentally, we show that SGM significantly outperforms current state of the art methods on long horizon, sparse-reward visual navigation tasks. Project video and code are available at this https URL" @default.
- W3010717779 created "2020-03-23" @default.
- W3010717779 creator A5005430934 @default.
- W3010717779 creator A5016023094 @default.
- W3010717779 creator A5049349154 @default.
- W3010717779 creator A5051005226 @default.
- W3010717779 creator A5066151318 @default.
- W3010717779 creator A5087643379 @default.
- W3010717779 date "2020-03-13" @default.
- W3010717779 modified "2023-09-27" @default.
- W3010717779 title "Sparse Graphical Memory for Robust Planning." @default.
- W3010717779 cites W131069610 @default.
- W3010717779 cites W1555801537 @default.
- W3010717779 cites W1577352482 @default.
- W3010717779 cites W1594201624 @default.
- W3010717779 cites W1757796397 @default.
- W3010717779 cites W1771410628 @default.
- W3010717779 cites W1959608418 @default.
- W3010717779 cites W1969483458 @default.
- W3010717779 cites W2000214310 @default.
- W3010717779 cites W2003159902 @default.
- W3010717779 cites W2058735307 @default.
- W3010717779 cites W2059453880 @default.
- W3010717779 cites W2073787051 @default.
- W3010717779 cites W2080132051 @default.
- W3010717779 cites W2122398932 @default.
- W3010717779 cites W2126677653 @default.
- W3010717779 cites W2127578024 @default.
- W3010717779 cites W2141911445 @default.
- W3010717779 cites W2143996311 @default.
- W3010717779 cites W2145339207 @default.
- W3010717779 cites W2146881125 @default.
- W3010717779 cites W2149698362 @default.
- W3010717779 cites W2169528473 @default.
- W3010717779 cites W2397240726 @default.
- W3010717779 cites W2557465155 @default.
- W3010717779 cites W2568646110 @default.
- W3010717779 cites W2736601468 @default.
- W3010717779 cites W2753738274 @default.
- W3010717779 cites W2798705390 @default.
- W3010717779 cites W2904246096 @default.
- W3010717779 cites W2918642789 @default.
- W3010717779 cites W2962812366 @default.
- W3010717779 cites W2963241167 @default.
- W3010717779 cites W2963245725 @default.
- W3010717779 cites W2963864421 @default.
- W3010717779 cites W2963948533 @default.
- W3010717779 cites W2964001908 @default.
- W3010717779 cites W2964036701 @default.
- W3010717779 cites W2964121744 @default.
- W3010717779 cites W2964342357 @default.
- W3010717779 cites W2970387978 @default.
- W3010717779 cites W2970720334 @default.
- W3010717779 cites W2997207934 @default.
- W3010717779 cites W3034728521 @default.
- W3010717779 cites W3091037708 @default.
- W3010717779 cites W567721252 @default.
- W3010717779 cites W2955993556 @default.
- W3010717779 hasPublicationYear "2020" @default.
- W3010717779 type Work @default.
- W3010717779 sameAs 3010717779 @default.
- W3010717779 citedByCount "0" @default.
- W3010717779 crossrefType "posted-content" @default.
- W3010717779 hasAuthorship W3010717779A5005430934 @default.
- W3010717779 hasAuthorship W3010717779A5016023094 @default.
- W3010717779 hasAuthorship W3010717779A5049349154 @default.
- W3010717779 hasAuthorship W3010717779A5051005226 @default.
- W3010717779 hasAuthorship W3010717779A5066151318 @default.
- W3010717779 hasAuthorship W3010717779A5087643379 @default.
- W3010717779 hasConcept C11413529 @default.
- W3010717779 hasConcept C126255220 @default.
- W3010717779 hasConcept C12713177 @default.
- W3010717779 hasConcept C154945302 @default.
- W3010717779 hasConcept C155846161 @default.
- W3010717779 hasConcept C177264268 @default.
- W3010717779 hasConcept C199360897 @default.
- W3010717779 hasConcept C2776436953 @default.
- W3010717779 hasConcept C2776760102 @default.
- W3010717779 hasConcept C28761237 @default.
- W3010717779 hasConcept C33923547 @default.
- W3010717779 hasConcept C41008148 @default.
- W3010717779 hasConcept C48103436 @default.
- W3010717779 hasConcept C80444323 @default.
- W3010717779 hasConcept C97541855 @default.
- W3010717779 hasConceptScore W3010717779C11413529 @default.
- W3010717779 hasConceptScore W3010717779C126255220 @default.
- W3010717779 hasConceptScore W3010717779C12713177 @default.
- W3010717779 hasConceptScore W3010717779C154945302 @default.
- W3010717779 hasConceptScore W3010717779C155846161 @default.
- W3010717779 hasConceptScore W3010717779C177264268 @default.
- W3010717779 hasConceptScore W3010717779C199360897 @default.
- W3010717779 hasConceptScore W3010717779C2776436953 @default.
- W3010717779 hasConceptScore W3010717779C2776760102 @default.
- W3010717779 hasConceptScore W3010717779C28761237 @default.
- W3010717779 hasConceptScore W3010717779C33923547 @default.
- W3010717779 hasConceptScore W3010717779C41008148 @default.
- W3010717779 hasConceptScore W3010717779C48103436 @default.
- W3010717779 hasConceptScore W3010717779C80444323 @default.
- W3010717779 hasConceptScore W3010717779C97541855 @default.
- W3010717779 hasLocation W30107177791 @default.