Matches in SemOpenAlex for { <https://semopenalex.org/work/W4387559701> ?p ?o ?g. }
Showing items 1 to 89 of
89
with 100 items per page.
- W4387559701 abstract "In visual-based Reinforcement Learning (RL), agents often struggle to generalize well to environmental variations in the state space that were not observed during training. The variations can arise in both task-irrelevant features, such as background noise, and task-relevant features, such as robot configurations, that are related to the optimal decisions. To achieve generalization in both situations, agents are required to accurately understand the impact of changed features on the decisions, i.e., establishing the true associations between changed features and decisions in the policy model. However, due to the inherent correlations among features in the state space, the associations between features and decisions become entangled, making it difficult for the policy to distinguish them. To this end, we propose Saliency-Guided Features Decorrelation (SGFD) to eliminate these correlations through sample reweighting. Concretely, SGFD consists of two core techniques: Random Fourier Functions (RFF) and the saliency map. RFF is utilized to estimate the complex non-linear correlations in high-dimensional images, while the saliency map is designed to identify the changed features. Under the guidance of the saliency map, SGFD employs sample reweighting to minimize the estimated correlations related to changed features, thereby achieving decorrelation in visual RL tasks. Our experimental results demonstrate that SGFD can generalize well on a wide range of test environments and significantly outperforms state-of-the-art methods in handling both task-irrelevant variations and task-relevant variations." @default.
- W4387559701 created "2023-10-12" @default.
- W4387559701 creator A5001147061 @default.
- W4387559701 creator A5028694889 @default.
- W4387559701 creator A5063197394 @default.
- W4387559701 creator A5065676828 @default.
- W4387559701 creator A5067750782 @default.
- W4387559701 creator A5087644493 @default.
- W4387559701 creator A5088612760 @default.
- W4387559701 creator A5091537321 @default.
- W4387559701 date "2023-10-08" @default.
- W4387559701 modified "2023-10-18" @default.
- W4387559701 title "Learning Generalizable Agents via Saliency-Guided Features Decorrelation" @default.
- W4387559701 doi "https://doi.org/10.48550/arxiv.2310.05086" @default.
- W4387559701 hasPublicationYear "2023" @default.
- W4387559701 type Work @default.
- W4387559701 citedByCount "0" @default.
- W4387559701 crossrefType "posted-content" @default.
- W4387559701 hasAuthorship W4387559701A5001147061 @default.
- W4387559701 hasAuthorship W4387559701A5028694889 @default.
- W4387559701 hasAuthorship W4387559701A5063197394 @default.
- W4387559701 hasAuthorship W4387559701A5065676828 @default.
- W4387559701 hasAuthorship W4387559701A5067750782 @default.
- W4387559701 hasAuthorship W4387559701A5087644493 @default.
- W4387559701 hasAuthorship W4387559701A5088612760 @default.
- W4387559701 hasAuthorship W4387559701A5091537321 @default.
- W4387559701 hasBestOaLocation W43875597011 @default.
- W4387559701 hasConcept C105795698 @default.
- W4387559701 hasConcept C115961682 @default.
- W4387559701 hasConcept C119857082 @default.
- W4387559701 hasConcept C134306372 @default.
- W4387559701 hasConcept C153180895 @default.
- W4387559701 hasConcept C154945302 @default.
- W4387559701 hasConcept C159985019 @default.
- W4387559701 hasConcept C162324750 @default.
- W4387559701 hasConcept C177148314 @default.
- W4387559701 hasConcept C177860922 @default.
- W4387559701 hasConcept C185592680 @default.
- W4387559701 hasConcept C187736073 @default.
- W4387559701 hasConcept C192562407 @default.
- W4387559701 hasConcept C198531522 @default.
- W4387559701 hasConcept C204323151 @default.
- W4387559701 hasConcept C2780451532 @default.
- W4387559701 hasConcept C31972630 @default.
- W4387559701 hasConcept C33923547 @default.
- W4387559701 hasConcept C41008148 @default.
- W4387559701 hasConcept C43617362 @default.
- W4387559701 hasConcept C72434380 @default.
- W4387559701 hasConcept C97541855 @default.
- W4387559701 hasConcept C99498987 @default.
- W4387559701 hasConceptScore W4387559701C105795698 @default.
- W4387559701 hasConceptScore W4387559701C115961682 @default.
- W4387559701 hasConceptScore W4387559701C119857082 @default.
- W4387559701 hasConceptScore W4387559701C134306372 @default.
- W4387559701 hasConceptScore W4387559701C153180895 @default.
- W4387559701 hasConceptScore W4387559701C154945302 @default.
- W4387559701 hasConceptScore W4387559701C159985019 @default.
- W4387559701 hasConceptScore W4387559701C162324750 @default.
- W4387559701 hasConceptScore W4387559701C177148314 @default.
- W4387559701 hasConceptScore W4387559701C177860922 @default.
- W4387559701 hasConceptScore W4387559701C185592680 @default.
- W4387559701 hasConceptScore W4387559701C187736073 @default.
- W4387559701 hasConceptScore W4387559701C192562407 @default.
- W4387559701 hasConceptScore W4387559701C198531522 @default.
- W4387559701 hasConceptScore W4387559701C204323151 @default.
- W4387559701 hasConceptScore W4387559701C2780451532 @default.
- W4387559701 hasConceptScore W4387559701C31972630 @default.
- W4387559701 hasConceptScore W4387559701C33923547 @default.
- W4387559701 hasConceptScore W4387559701C41008148 @default.
- W4387559701 hasConceptScore W4387559701C43617362 @default.
- W4387559701 hasConceptScore W4387559701C72434380 @default.
- W4387559701 hasConceptScore W4387559701C97541855 @default.
- W4387559701 hasConceptScore W4387559701C99498987 @default.
- W4387559701 hasLocation W43875597011 @default.
- W4387559701 hasOpenAccess W4387559701 @default.
- W4387559701 hasPrimaryLocation W43875597011 @default.
- W4387559701 hasRelatedWork W1604939135 @default.
- W4387559701 hasRelatedWork W2126211886 @default.
- W4387559701 hasRelatedWork W2350784623 @default.
- W4387559701 hasRelatedWork W2992629954 @default.
- W4387559701 hasRelatedWork W2999580272 @default.
- W4387559701 hasRelatedWork W3009457412 @default.
- W4387559701 hasRelatedWork W3212257828 @default.
- W4387559701 hasRelatedWork W4297873223 @default.
- W4387559701 hasRelatedWork W4377293004 @default.
- W4387559701 hasRelatedWork W4225571923 @default.
- W4387559701 isParatext "false" @default.
- W4387559701 isRetracted "false" @default.
- W4387559701 workType "article" @default.