Matches in SemOpenAlex for { <https://semopenalex.org/work/W4226403986> ?p ?o ?g. }
- W4226403986 abstract "Reinforcement learning can train policies that effectively perform complex tasks. However for long-horizon tasks, the performance of these methods degrades with horizon, often necessitating reasoning over and chaining lower-level skills. Hierarchical reinforcement learning aims to enable this by providing a bank of low-level skills as action abstractions. Hierarchies can further improve on this by abstracting the space states as well. We posit that a suitable state abstraction should depend on the capabilities of the available lower-level policies. We propose Value Function Spaces: a simple approach that produces such a representation by using the value functions corresponding to each lower-level skill. These value functions capture the affordances of the scene, thus forming a representation that compactly abstracts task relevant information and robustly ignores distractors. Empirical evaluations for maze-solving and robotic manipulation tasks demonstrate that our approach improves long-horizon performance and enables better zero-shot generalization than alternative model-free and model-based methods." @default.
- W4226403986 created "2022-05-05" @default.
- W4226403986 creator A5011678003 @default.
- W4226403986 creator A5018507768 @default.
- W4226403986 creator A5025016495 @default.
- W4226403986 creator A5026322200 @default.
- W4226403986 creator A5044385533 @default.
- W4226403986 creator A5049184232 @default.
- W4226403986 creator A5064265174 @default.
- W4226403986 date "2021-11-04" @default.
- W4226403986 modified "2023-09-24" @default.
- W4226403986 title "Value Function Spaces: Skill-Centric State Abstractions for Long-Horizon Reasoning" @default.
- W4226403986 doi "https://doi.org/10.48550/arxiv.2111.03189" @default.
- W4226403986 hasPublicationYear "2021" @default.
- W4226403986 type Work @default.
- W4226403986 citedByCount "0" @default.
- W4226403986 crossrefType "posted-content" @default.
- W4226403986 hasAuthorship W4226403986A5011678003 @default.
- W4226403986 hasAuthorship W4226403986A5018507768 @default.
- W4226403986 hasAuthorship W4226403986A5025016495 @default.
- W4226403986 hasAuthorship W4226403986A5026322200 @default.
- W4226403986 hasAuthorship W4226403986A5044385533 @default.
- W4226403986 hasAuthorship W4226403986A5049184232 @default.
- W4226403986 hasAuthorship W4226403986A5064265174 @default.
- W4226403986 hasBestOaLocation W42264039861 @default.
- W4226403986 hasConcept C105795698 @default.
- W4226403986 hasConcept C107457646 @default.
- W4226403986 hasConcept C111472728 @default.
- W4226403986 hasConcept C119857082 @default.
- W4226403986 hasConcept C124304363 @default.
- W4226403986 hasConcept C126255220 @default.
- W4226403986 hasConcept C134306372 @default.
- W4226403986 hasConcept C138885662 @default.
- W4226403986 hasConcept C14036430 @default.
- W4226403986 hasConcept C14646407 @default.
- W4226403986 hasConcept C154945302 @default.
- W4226403986 hasConcept C15744967 @default.
- W4226403986 hasConcept C159176650 @default.
- W4226403986 hasConcept C162324750 @default.
- W4226403986 hasConcept C177148314 @default.
- W4226403986 hasConcept C17744445 @default.
- W4226403986 hasConcept C187736073 @default.
- W4226403986 hasConcept C194995250 @default.
- W4226403986 hasConcept C199539241 @default.
- W4226403986 hasConcept C2524010 @default.
- W4226403986 hasConcept C2776291640 @default.
- W4226403986 hasConcept C2776359362 @default.
- W4226403986 hasConcept C2780451532 @default.
- W4226403986 hasConcept C33923547 @default.
- W4226403986 hasConcept C41008148 @default.
- W4226403986 hasConcept C49020025 @default.
- W4226403986 hasConcept C542102704 @default.
- W4226403986 hasConcept C72434380 @default.
- W4226403986 hasConcept C78458016 @default.
- W4226403986 hasConcept C86803240 @default.
- W4226403986 hasConcept C94625758 @default.
- W4226403986 hasConcept C97541855 @default.
- W4226403986 hasConceptScore W4226403986C105795698 @default.
- W4226403986 hasConceptScore W4226403986C107457646 @default.
- W4226403986 hasConceptScore W4226403986C111472728 @default.
- W4226403986 hasConceptScore W4226403986C119857082 @default.
- W4226403986 hasConceptScore W4226403986C124304363 @default.
- W4226403986 hasConceptScore W4226403986C126255220 @default.
- W4226403986 hasConceptScore W4226403986C134306372 @default.
- W4226403986 hasConceptScore W4226403986C138885662 @default.
- W4226403986 hasConceptScore W4226403986C14036430 @default.
- W4226403986 hasConceptScore W4226403986C14646407 @default.
- W4226403986 hasConceptScore W4226403986C154945302 @default.
- W4226403986 hasConceptScore W4226403986C15744967 @default.
- W4226403986 hasConceptScore W4226403986C159176650 @default.
- W4226403986 hasConceptScore W4226403986C162324750 @default.
- W4226403986 hasConceptScore W4226403986C177148314 @default.
- W4226403986 hasConceptScore W4226403986C17744445 @default.
- W4226403986 hasConceptScore W4226403986C187736073 @default.
- W4226403986 hasConceptScore W4226403986C194995250 @default.
- W4226403986 hasConceptScore W4226403986C199539241 @default.
- W4226403986 hasConceptScore W4226403986C2524010 @default.
- W4226403986 hasConceptScore W4226403986C2776291640 @default.
- W4226403986 hasConceptScore W4226403986C2776359362 @default.
- W4226403986 hasConceptScore W4226403986C2780451532 @default.
- W4226403986 hasConceptScore W4226403986C33923547 @default.
- W4226403986 hasConceptScore W4226403986C41008148 @default.
- W4226403986 hasConceptScore W4226403986C49020025 @default.
- W4226403986 hasConceptScore W4226403986C542102704 @default.
- W4226403986 hasConceptScore W4226403986C72434380 @default.
- W4226403986 hasConceptScore W4226403986C78458016 @default.
- W4226403986 hasConceptScore W4226403986C86803240 @default.
- W4226403986 hasConceptScore W4226403986C94625758 @default.
- W4226403986 hasConceptScore W4226403986C97541855 @default.
- W4226403986 hasLocation W42264039861 @default.
- W4226403986 hasOpenAccess W4226403986 @default.
- W4226403986 hasPrimaryLocation W42264039861 @default.
- W4226403986 hasRelatedWork W2621131026 @default.
- W4226403986 hasRelatedWork W2764311431 @default.
- W4226403986 hasRelatedWork W2789410350 @default.
- W4226403986 hasRelatedWork W3166968915 @default.
- W4226403986 hasRelatedWork W4206848870 @default.
- W4226403986 hasRelatedWork W4283800590 @default.
- W4226403986 hasRelatedWork W4297789760 @default.
- W4226403986 hasRelatedWork W4302011254 @default.