Matches in SemOpenAlex for { <https://semopenalex.org/work/W3213373963> ?p ?o ?g. }
- W3213373963 abstract "Abstract Algorithmic design in neural architecture search (NAS) has received a lot of attention, aiming to improve performance and reduce computational cost. Despite the great advances made, few authors have proposed to tailor initialization techniques for NAS. However, the literature shows that a good initial set of solutions facilitates finding the optima. Therefore, in this study, we propose a data-driven technique to initialize a population-based NAS algorithm. First, we perform a calibrated clustering analysis of the search space, and second, we extract the centroids and use them to initialize a NAS algorithm. We benchmark our proposed approach against random and Latin hypercube sampling initialization using three population-based algorithms, namely a genetic algorithm, an evolutionary algorithm, and aging evolution, on CIFAR-10. More specifically, we use NAS-Bench-101 to leverage the availability of NAS benchmarks. The results show that compared to random and Latin hypercube sampling, the proposed initialization technique enables achieving significant long-term improvements for two of the search baselines, and sometimes in various search scenarios (various training budget). Besides, we also investigate how an initial population gathered on the tabular benchmark can be used for improving search on another dataset, the So2Sat LCZ-42. Our results show similar improvements on the target dataset, despite a limited training budget. Moreover, we analyse the distributions of solutions obtained and find that that the population provided by the data-driven initialization technique enables retrieving local optima (maxima) of high fitness and similar configurations." @default.
- W3213373963 created "2021-11-22" @default.
- W3213373963 creator A5002546446 @default.
- W3213373963 creator A5068384981 @default.
- W3213373963 creator A5090258132 @default.
- W3213373963 date "2023-03-22" @default.
- W3213373963 modified "2023-09-26" @default.
- W3213373963 title "A data-driven approach to neural architecture search initialization" @default.
- W3213373963 cites W1502203485 @default.
- W3213373963 cites W1508425855 @default.
- W3213373963 cites W1559956479 @default.
- W3213373963 cites W1571031121 @default.
- W3213373963 cites W1769038400 @default.
- W3213373963 cites W1987971958 @default.
- W3213373963 cites W2002592134 @default.
- W3213373963 cites W2003176983 @default.
- W3213373963 cites W2051224630 @default.
- W3213373963 cites W2068696225 @default.
- W3213373963 cites W2072245049 @default.
- W3213373963 cites W2082754320 @default.
- W3213373963 cites W2085487226 @default.
- W3213373963 cites W2093872789 @default.
- W3213373963 cites W2111935653 @default.
- W3213373963 cites W2128898602 @default.
- W3213373963 cites W2157833270 @default.
- W3213373963 cites W2254028986 @default.
- W3213373963 cites W2394896783 @default.
- W3213373963 cites W2403365423 @default.
- W3213373963 cites W2587345921 @default.
- W3213373963 cites W2593744649 @default.
- W3213373963 cites W2892599100 @default.
- W3213373963 cites W2919115771 @default.
- W3213373963 cites W2959643294 @default.
- W3213373963 cites W2963029020 @default.
- W3213373963 cites W2964121818 @default.
- W3213373963 cites W2965658867 @default.
- W3213373963 cites W2975160087 @default.
- W3213373963 cites W2977748703 @default.
- W3213373963 cites W3007766931 @default.
- W3213373963 cites W3107453328 @default.
- W3213373963 cites W3131407372 @default.
- W3213373963 cites W3139552059 @default.
- W3213373963 cites W3163767593 @default.
- W3213373963 cites W3174844646 @default.
- W3213373963 cites W4213308398 @default.
- W3213373963 cites W4214513763 @default.
- W3213373963 cites W61959169 @default.
- W3213373963 doi "https://doi.org/10.1007/s10472-022-09823-0" @default.
- W3213373963 hasPublicationYear "2023" @default.
- W3213373963 type Work @default.
- W3213373963 sameAs 3213373963 @default.
- W3213373963 citedByCount "0" @default.
- W3213373963 crossrefType "journal-article" @default.
- W3213373963 hasAuthorship W3213373963A5002546446 @default.
- W3213373963 hasAuthorship W3213373963A5068384981 @default.
- W3213373963 hasAuthorship W3213373963A5090258132 @default.
- W3213373963 hasBestOaLocation W32133739631 @default.
- W3213373963 hasConcept C105795698 @default.
- W3213373963 hasConcept C114466953 @default.
- W3213373963 hasConcept C119857082 @default.
- W3213373963 hasConcept C124101348 @default.
- W3213373963 hasConcept C126255220 @default.
- W3213373963 hasConcept C13280743 @default.
- W3213373963 hasConcept C135320971 @default.
- W3213373963 hasConcept C141934464 @default.
- W3213373963 hasConcept C144024400 @default.
- W3213373963 hasConcept C149923435 @default.
- W3213373963 hasConcept C153083717 @default.
- W3213373963 hasConcept C154945302 @default.
- W3213373963 hasConcept C159149176 @default.
- W3213373963 hasConcept C185798385 @default.
- W3213373963 hasConcept C19499675 @default.
- W3213373963 hasConcept C199360897 @default.
- W3213373963 hasConcept C205649164 @default.
- W3213373963 hasConcept C20820323 @default.
- W3213373963 hasConcept C2908647359 @default.
- W3213373963 hasConcept C33923547 @default.
- W3213373963 hasConcept C41008148 @default.
- W3213373963 hasConcept C73555534 @default.
- W3213373963 hasConceptScore W3213373963C105795698 @default.
- W3213373963 hasConceptScore W3213373963C114466953 @default.
- W3213373963 hasConceptScore W3213373963C119857082 @default.
- W3213373963 hasConceptScore W3213373963C124101348 @default.
- W3213373963 hasConceptScore W3213373963C126255220 @default.
- W3213373963 hasConceptScore W3213373963C13280743 @default.
- W3213373963 hasConceptScore W3213373963C135320971 @default.
- W3213373963 hasConceptScore W3213373963C141934464 @default.
- W3213373963 hasConceptScore W3213373963C144024400 @default.
- W3213373963 hasConceptScore W3213373963C149923435 @default.
- W3213373963 hasConceptScore W3213373963C153083717 @default.
- W3213373963 hasConceptScore W3213373963C154945302 @default.
- W3213373963 hasConceptScore W3213373963C159149176 @default.
- W3213373963 hasConceptScore W3213373963C185798385 @default.
- W3213373963 hasConceptScore W3213373963C19499675 @default.
- W3213373963 hasConceptScore W3213373963C199360897 @default.
- W3213373963 hasConceptScore W3213373963C205649164 @default.
- W3213373963 hasConceptScore W3213373963C20820323 @default.
- W3213373963 hasConceptScore W3213373963C2908647359 @default.
- W3213373963 hasConceptScore W3213373963C33923547 @default.
- W3213373963 hasConceptScore W3213373963C41008148 @default.