Matches in SemOpenAlex for { <https://semopenalex.org/work/W3119292076> ?p ?o ?g. }
Showing items 1 to 90 of
90
with 100 items per page.
- W3119292076 abstract "Existing visual navigation approaches leverage classification neural networks to extract global features from visual data for navigation. However, these networks are not originally designed for navigation tasks. Thus, the neural architectures might not be suitable to capture scene con-tents. Fortunately, neural architecture search (NAS) brings a hope to solve this problem. In this paper, we propose an Auto-Navigator to customize a specialized network for visual navigation. However, as navigation tasks mainly rely on reinforcement learning (RL) rewards in training, such weak supervision is insufficiently indicative for NAS to optimize visual perception network. Thus, we introduce imitation learning (IL) with optimal paths to optimize navigation policies while selecting an optimal architecture. As Auto-Navigator can obtain a direct supervision in every step, such guidance greatly facilitates architecture search. In particular, we initialize our Auto-Navigator with a learnable distribution over the search space of visual perception architecture, and then optimize the distribution with IL supervision. Afterwards, we employ an RL reward function to fine-tune our Auto-Navigator to improve the generalization ability of our model. Extensive experiments demonstrate that our Auto-Navigator outperforms baseline methods on Gibson and Matterport3D without significantly increasing network parameters." @default.
- W3119292076 created "2021-01-18" @default.
- W3119292076 creator A5003076238 @default.
- W3119292076 creator A5005421447 @default.
- W3119292076 creator A5068008058 @default.
- W3119292076 creator A5069837273 @default.
- W3119292076 date "2021-01-01" @default.
- W3119292076 modified "2023-10-18" @default.
- W3119292076 title "Auto-Navigator: Decoupled Neural Architecture Search for Visual Navigation" @default.
- W3119292076 cites W1424654272 @default.
- W3119292076 cites W1986014385 @default.
- W3119292076 cites W2020890121 @default.
- W3119292076 cites W2124386111 @default.
- W3119292076 cites W2146881125 @default.
- W3119292076 cites W2194775991 @default.
- W3119292076 cites W2593841437 @default.
- W3119292076 cites W2960010704 @default.
- W3119292076 cites W2962732398 @default.
- W3119292076 cites W2962744691 @default.
- W3119292076 cites W2962789679 @default.
- W3119292076 cites W2962887844 @default.
- W3119292076 cites W2964081807 @default.
- W3119292076 cites W2965658867 @default.
- W3119292076 cites W2965949912 @default.
- W3119292076 cites W2967733054 @default.
- W3119292076 cites W2967853831 @default.
- W3119292076 cites W2973665503 @default.
- W3119292076 cites W2980270353 @default.
- W3119292076 cites W2981836239 @default.
- W3119292076 cites W2983335573 @default.
- W3119292076 cites W2986093954 @default.
- W3119292076 cites W2986357018 @default.
- W3119292076 cites W2998025051 @default.
- W3119292076 cites W3109001458 @default.
- W3119292076 cites W3124420883 @default.
- W3119292076 doi "https://doi.org/10.1109/wacv48630.2021.00379" @default.
- W3119292076 hasPublicationYear "2021" @default.
- W3119292076 type Work @default.
- W3119292076 sameAs 3119292076 @default.
- W3119292076 citedByCount "5" @default.
- W3119292076 countsByYear W31192920762021 @default.
- W3119292076 countsByYear W31192920762022 @default.
- W3119292076 crossrefType "proceedings-article" @default.
- W3119292076 hasAuthorship W3119292076A5003076238 @default.
- W3119292076 hasAuthorship W3119292076A5005421447 @default.
- W3119292076 hasAuthorship W3119292076A5068008058 @default.
- W3119292076 hasAuthorship W3119292076A5069837273 @default.
- W3119292076 hasConcept C119857082 @default.
- W3119292076 hasConcept C123657996 @default.
- W3119292076 hasConcept C142362112 @default.
- W3119292076 hasConcept C153083717 @default.
- W3119292076 hasConcept C153349607 @default.
- W3119292076 hasConcept C154945302 @default.
- W3119292076 hasConcept C158495155 @default.
- W3119292076 hasConcept C193415008 @default.
- W3119292076 hasConcept C31972630 @default.
- W3119292076 hasConcept C38652104 @default.
- W3119292076 hasConcept C41008148 @default.
- W3119292076 hasConcept C50644808 @default.
- W3119292076 hasConcept C97541855 @default.
- W3119292076 hasConceptScore W3119292076C119857082 @default.
- W3119292076 hasConceptScore W3119292076C123657996 @default.
- W3119292076 hasConceptScore W3119292076C142362112 @default.
- W3119292076 hasConceptScore W3119292076C153083717 @default.
- W3119292076 hasConceptScore W3119292076C153349607 @default.
- W3119292076 hasConceptScore W3119292076C154945302 @default.
- W3119292076 hasConceptScore W3119292076C158495155 @default.
- W3119292076 hasConceptScore W3119292076C193415008 @default.
- W3119292076 hasConceptScore W3119292076C31972630 @default.
- W3119292076 hasConceptScore W3119292076C38652104 @default.
- W3119292076 hasConceptScore W3119292076C41008148 @default.
- W3119292076 hasConceptScore W3119292076C50644808 @default.
- W3119292076 hasConceptScore W3119292076C97541855 @default.
- W3119292076 hasLocation W31192920761 @default.
- W3119292076 hasOpenAccess W3119292076 @default.
- W3119292076 hasPrimaryLocation W31192920761 @default.
- W3119292076 hasRelatedWork W1891287906 @default.
- W3119292076 hasRelatedWork W260766989 @default.
- W3119292076 hasRelatedWork W2959276766 @default.
- W3119292076 hasRelatedWork W2961085424 @default.
- W3119292076 hasRelatedWork W3037422413 @default.
- W3119292076 hasRelatedWork W3170628611 @default.
- W3119292076 hasRelatedWork W4206669594 @default.
- W3119292076 hasRelatedWork W4295941380 @default.
- W3119292076 hasRelatedWork W4297949354 @default.
- W3119292076 hasRelatedWork W4319083788 @default.
- W3119292076 isParatext "false" @default.
- W3119292076 isRetracted "false" @default.
- W3119292076 magId "3119292076" @default.
- W3119292076 workType "article" @default.