Matches in SemOpenAlex for { <https://semopenalex.org/work/W4387311340> ?p ?o ?g. }
Showing items 1 to 80 of
80
with 100 items per page.
- W4387311340 endingPage "121742" @default.
- W4387311340 startingPage "121742" @default.
- W4387311340 abstract "Predictor-based Neural Architecture Search is an important topic since it can efficiently reduce the computational cost of evaluating candidate architectures. Most existing predictor-based NAS algorithms aim to design different predictors to improve prediction performance. Unfortunately, even a promising performance predictor may suffer from the accuracy decline due to long-term and continuous usage, thus leading to the degraded performance of the search strategy. That naturally gives rise to the following problems: how do predictors affect search strategies and how to efficiently use the predictor? In this paper, we take Reinforcement Learning (RL) based search strategy to study theoretically and empirically the impact of predictors on search strategies. We first formulate an RL-Predictor-based NAS algorithm as model-based RL and analyze it with a guarantee of monotonic improvement. Then, based on this analysis, we propose a simple procedure of predictor usage, named mixedbatch, which contains ground-truth data and prediction data in a batch. The proposed procedure can efficiently reduce the impact of predictor errors on the RL-based search strategy with maintaining performance growth. Our algorithm, RL-Predictor-based Neural Architecture Search with Mixed batch (RPNASM), outperforms traditional NAS algorithms and prior state-of-the-art predictor-based NAS algorithms on three NAS-Bench-201 tasks and one NAS-Bench-ASR task. Our code is available at https://github.com/tjdeng/RPNASM." @default.
- W4387311340 created "2023-10-04" @default.
- W4387311340 creator A5010197223 @default.
- W4387311340 creator A5011881849 @default.
- W4387311340 creator A5088941117 @default.
- W4387311340 date "2024-03-01" @default.
- W4387311340 modified "2023-10-17" @default.
- W4387311340 title "How predictors affect the RL-based search strategy in Neural Architecture Search?" @default.
- W4387311340 cites W2102928860 @default.
- W4387311340 cites W2107726111 @default.
- W4387311340 cites W2119717200 @default.
- W4387311340 cites W2963821229 @default.
- W4387311340 cites W2964081807 @default.
- W4387311340 cites W2965658867 @default.
- W4387311340 cites W2981406437 @default.
- W4387311340 cites W3013310686 @default.
- W4387311340 cites W3034723893 @default.
- W4387311340 cites W3035715446 @default.
- W4387311340 cites W3082154327 @default.
- W4387311340 cites W3102476541 @default.
- W4387311340 cites W3173166478 @default.
- W4387311340 cites W4300772200 @default.
- W4387311340 cites W4382317589 @default.
- W4387311340 doi "https://doi.org/10.1016/j.eswa.2023.121742" @default.
- W4387311340 hasPublicationYear "2024" @default.
- W4387311340 type Work @default.
- W4387311340 citedByCount "0" @default.
- W4387311340 crossrefType "journal-article" @default.
- W4387311340 hasAuthorship W4387311340A5010197223 @default.
- W4387311340 hasAuthorship W4387311340A5011881849 @default.
- W4387311340 hasAuthorship W4387311340A5088941117 @default.
- W4387311340 hasBestOaLocation W43873113401 @default.
- W4387311340 hasConcept C11413529 @default.
- W4387311340 hasConcept C119857082 @default.
- W4387311340 hasConcept C124101348 @default.
- W4387311340 hasConcept C125583679 @default.
- W4387311340 hasConcept C154945302 @default.
- W4387311340 hasConcept C162324750 @default.
- W4387311340 hasConcept C177264268 @default.
- W4387311340 hasConcept C187736073 @default.
- W4387311340 hasConcept C199360897 @default.
- W4387311340 hasConcept C2776760102 @default.
- W4387311340 hasConcept C2780451532 @default.
- W4387311340 hasConcept C41008148 @default.
- W4387311340 hasConcept C50644808 @default.
- W4387311340 hasConcept C97541855 @default.
- W4387311340 hasConceptScore W4387311340C11413529 @default.
- W4387311340 hasConceptScore W4387311340C119857082 @default.
- W4387311340 hasConceptScore W4387311340C124101348 @default.
- W4387311340 hasConceptScore W4387311340C125583679 @default.
- W4387311340 hasConceptScore W4387311340C154945302 @default.
- W4387311340 hasConceptScore W4387311340C162324750 @default.
- W4387311340 hasConceptScore W4387311340C177264268 @default.
- W4387311340 hasConceptScore W4387311340C187736073 @default.
- W4387311340 hasConceptScore W4387311340C199360897 @default.
- W4387311340 hasConceptScore W4387311340C2776760102 @default.
- W4387311340 hasConceptScore W4387311340C2780451532 @default.
- W4387311340 hasConceptScore W4387311340C41008148 @default.
- W4387311340 hasConceptScore W4387311340C50644808 @default.
- W4387311340 hasConceptScore W4387311340C97541855 @default.
- W4387311340 hasFunder F4320321001 @default.
- W4387311340 hasLocation W43873113401 @default.
- W4387311340 hasOpenAccess W4387311340 @default.
- W4387311340 hasPrimaryLocation W43873113401 @default.
- W4387311340 hasRelatedWork W2024136090 @default.
- W4387311340 hasRelatedWork W2031695474 @default.
- W4387311340 hasRelatedWork W2138720691 @default.
- W4387311340 hasRelatedWork W2586732548 @default.
- W4387311340 hasRelatedWork W2964765435 @default.
- W4387311340 hasRelatedWork W3049728571 @default.
- W4387311340 hasRelatedWork W4306904969 @default.
- W4387311340 hasRelatedWork W4362501864 @default.
- W4387311340 hasRelatedWork W4380318855 @default.
- W4387311340 hasRelatedWork W2585069576 @default.
- W4387311340 hasVolume "238" @default.
- W4387311340 isParatext "false" @default.
- W4387311340 isRetracted "false" @default.
- W4387311340 workType "article" @default.