Matches in SemOpenAlex for { <https://semopenalex.org/work/W2897328500> ?p ?o ?g. }
- W2897328500 abstract "The smallest eigenvectors of the graph Laplacian are well-known to provide a succinct representation of the geometry of a weighted graph. In reinforcement learning (RL), where the weighted graph may be interpreted as the state transition process induced by a behavior policy acting on the environment, approximating the eigenvectors of the Laplacian provides a promising approach to state representation learning. However, existing methods for performing this approximation are ill-suited in general RL settings for two main reasons: First, they are computationally expensive, often requiring operations on large matrices. Second, these methods lack adequate justification beyond simple, tabular, finite-state settings. In this paper, we present a fully general and scalable method for approximating the eigenvectors of the Laplacian in a model-free RL context. We systematically evaluate our approach and empirically show that it generalizes beyond the tabular, finite-state setting. Even in tabular, finite-state settings, its ability to approximate the eigenvectors outperforms previous proposals. Finally, we show the potential benefits of using a Laplacian representation learned using our method in goal-achieving RL tasks, providing evidence that our technique can be used to significantly improve the performance of an RL agent." @default.
- W2897328500 created "2018-10-26" @default.
- W2897328500 creator A5020044791 @default.
- W2897328500 creator A5048032272 @default.
- W2897328500 creator A5057773393 @default.
- W2897328500 date "2018-10-10" @default.
- W2897328500 modified "2023-09-27" @default.
- W2897328500 title "The Laplacian in RL: Learning Representations with Efficient Approximations" @default.
- W2897328500 cites W1530948733 @default.
- W2897328500 cites W1580095995 @default.
- W2897328500 cites W1590252182 @default.
- W2897328500 cites W1904406446 @default.
- W2897328500 cites W2050583479 @default.
- W2897328500 cites W2143958939 @default.
- W2897328500 cites W2163176541 @default.
- W2897328500 cites W2165874743 @default.
- W2897328500 cites W2170239260 @default.
- W2897328500 cites W2561776174 @default.
- W2897328500 cites W2614839826 @default.
- W2897328500 cites W2765308067 @default.
- W2897328500 cites W2781711557 @default.
- W2897328500 cites W2787757704 @default.
- W2897328500 cites W2951605557 @default.
- W2897328500 hasPublicationYear "2018" @default.
- W2897328500 type Work @default.
- W2897328500 sameAs 2897328500 @default.
- W2897328500 citedByCount "4" @default.
- W2897328500 countsByYear W28973285002020 @default.
- W2897328500 countsByYear W28973285002021 @default.
- W2897328500 crossrefType "posted-content" @default.
- W2897328500 hasAuthorship W2897328500A5020044791 @default.
- W2897328500 hasAuthorship W2897328500A5048032272 @default.
- W2897328500 hasAuthorship W2897328500A5057773393 @default.
- W2897328500 hasConcept C11413529 @default.
- W2897328500 hasConcept C115178988 @default.
- W2897328500 hasConcept C119857082 @default.
- W2897328500 hasConcept C121332964 @default.
- W2897328500 hasConcept C126255220 @default.
- W2897328500 hasConcept C132525143 @default.
- W2897328500 hasConcept C134306372 @default.
- W2897328500 hasConcept C151730666 @default.
- W2897328500 hasConcept C154945302 @default.
- W2897328500 hasConcept C158693339 @default.
- W2897328500 hasConcept C165700671 @default.
- W2897328500 hasConcept C17744445 @default.
- W2897328500 hasConcept C199539241 @default.
- W2897328500 hasConcept C2776359362 @default.
- W2897328500 hasConcept C2779343474 @default.
- W2897328500 hasConcept C28826006 @default.
- W2897328500 hasConcept C2983497884 @default.
- W2897328500 hasConcept C33923547 @default.
- W2897328500 hasConcept C41008148 @default.
- W2897328500 hasConcept C48044578 @default.
- W2897328500 hasConcept C48103436 @default.
- W2897328500 hasConcept C62520636 @default.
- W2897328500 hasConcept C77088390 @default.
- W2897328500 hasConcept C80444323 @default.
- W2897328500 hasConcept C86803240 @default.
- W2897328500 hasConcept C94625758 @default.
- W2897328500 hasConcept C97541855 @default.
- W2897328500 hasConcept C98763669 @default.
- W2897328500 hasConceptScore W2897328500C11413529 @default.
- W2897328500 hasConceptScore W2897328500C115178988 @default.
- W2897328500 hasConceptScore W2897328500C119857082 @default.
- W2897328500 hasConceptScore W2897328500C121332964 @default.
- W2897328500 hasConceptScore W2897328500C126255220 @default.
- W2897328500 hasConceptScore W2897328500C132525143 @default.
- W2897328500 hasConceptScore W2897328500C134306372 @default.
- W2897328500 hasConceptScore W2897328500C151730666 @default.
- W2897328500 hasConceptScore W2897328500C154945302 @default.
- W2897328500 hasConceptScore W2897328500C158693339 @default.
- W2897328500 hasConceptScore W2897328500C165700671 @default.
- W2897328500 hasConceptScore W2897328500C17744445 @default.
- W2897328500 hasConceptScore W2897328500C199539241 @default.
- W2897328500 hasConceptScore W2897328500C2776359362 @default.
- W2897328500 hasConceptScore W2897328500C2779343474 @default.
- W2897328500 hasConceptScore W2897328500C28826006 @default.
- W2897328500 hasConceptScore W2897328500C2983497884 @default.
- W2897328500 hasConceptScore W2897328500C33923547 @default.
- W2897328500 hasConceptScore W2897328500C41008148 @default.
- W2897328500 hasConceptScore W2897328500C48044578 @default.
- W2897328500 hasConceptScore W2897328500C48103436 @default.
- W2897328500 hasConceptScore W2897328500C62520636 @default.
- W2897328500 hasConceptScore W2897328500C77088390 @default.
- W2897328500 hasConceptScore W2897328500C80444323 @default.
- W2897328500 hasConceptScore W2897328500C86803240 @default.
- W2897328500 hasConceptScore W2897328500C94625758 @default.
- W2897328500 hasConceptScore W2897328500C97541855 @default.
- W2897328500 hasConceptScore W2897328500C98763669 @default.
- W2897328500 hasLocation W28973285001 @default.
- W2897328500 hasOpenAccess W2897328500 @default.
- W2897328500 hasPrimaryLocation W28973285001 @default.
- W2897328500 hasRelatedWork W15354828 @default.
- W2897328500 hasRelatedWork W2097575529 @default.
- W2897328500 hasRelatedWork W2186878252 @default.
- W2897328500 hasRelatedWork W2225808822 @default.
- W2897328500 hasRelatedWork W2540616960 @default.
- W2897328500 hasRelatedWork W2560860718 @default.
- W2897328500 hasRelatedWork W2575860482 @default.
- W2897328500 hasRelatedWork W2809358500 @default.