Matches in SemOpenAlex for { <https://semopenalex.org/work/W3035888673> ?p ?o ?g. }
Showing items 1 to 90 of
90
with 100 items per page.
- W3035888673 abstract "With the advent of deeper, larger and more complex convolutional neural networks (CNN), manual design has become a daunting task, especially when hardware performance must be optimized. Sequential model-based optimization (SMBO) is an efficient method for hyperparameter optimization on highly parameterized machine learning (ML) algorithms, able to find good configurations with a limited number of evaluations by predicting the performance of candidates before evaluation. A case study on MNIST shows that SMBO regression model prediction error significantly impedes search performance in multi-objective optimization. To address this issue, we propose probabilistic SMBO, which selects candidates based on probabilistic estimation of their Pareto efficiency. With a formulation that incorporates error in accuracy prediction and uncertainty in latency measurement, probabilistic Pareto efficiency quantifies a candidate's quality in two ways: its likelihood of being Pareto optimal, and the expected number of current Pareto optimal solutions that it will dominate. We evaluate our proposed method on four image classification problems. Compared to a deterministic approach, probabilistic SMBO consistently generates Pareto optimal solutions that perform better, and that are competitive with state-of-the-art efficient CNN models, offering tremendous speedup in inference latency while maintaining comparable accuracy." @default.
- W3035888673 created "2020-06-25" @default.
- W3035888673 creator A5029544583 @default.
- W3035888673 creator A5034701871 @default.
- W3035888673 creator A5091600002 @default.
- W3035888673 date "2020-03-01" @default.
- W3035888673 modified "2023-10-16" @default.
- W3035888673 title "Probabilistic Sequential Multi-Objective Optimization of Convolutional Neural Networks" @default.
- W3035888673 cites W1558919105 @default.
- W3035888673 cites W2097117768 @default.
- W3035888673 cites W2106334424 @default.
- W3035888673 cites W2112796928 @default.
- W3035888673 cites W2117876524 @default.
- W3035888673 cites W2194775991 @default.
- W3035888673 cites W2534788641 @default.
- W3035888673 cites W2773706593 @default.
- W3035888673 cites W2808938483 @default.
- W3035888673 cites W2888850958 @default.
- W3035888673 cites W2890590696 @default.
- W3035888673 cites W2963163009 @default.
- W3035888673 cites W2963240979 @default.
- W3035888673 cites W2963446712 @default.
- W3035888673 cites W2963821229 @default.
- W3035888673 cites W2963918968 @default.
- W3035888673 cites W2964081807 @default.
- W3035888673 cites W60686164 @default.
- W3035888673 doi "https://doi.org/10.23919/date48585.2020.9116535" @default.
- W3035888673 hasPublicationYear "2020" @default.
- W3035888673 type Work @default.
- W3035888673 sameAs 3035888673 @default.
- W3035888673 citedByCount "3" @default.
- W3035888673 countsByYear W30358886732021 @default.
- W3035888673 countsByYear W30358886732022 @default.
- W3035888673 crossrefType "proceedings-article" @default.
- W3035888673 hasAuthorship W3035888673A5029544583 @default.
- W3035888673 hasAuthorship W3035888673A5034701871 @default.
- W3035888673 hasAuthorship W3035888673A5091600002 @default.
- W3035888673 hasConcept C111919701 @default.
- W3035888673 hasConcept C11413529 @default.
- W3035888673 hasConcept C119857082 @default.
- W3035888673 hasConcept C126255220 @default.
- W3035888673 hasConcept C137635306 @default.
- W3035888673 hasConcept C154945302 @default.
- W3035888673 hasConcept C165464430 @default.
- W3035888673 hasConcept C190502265 @default.
- W3035888673 hasConcept C2776214188 @default.
- W3035888673 hasConcept C2778049539 @default.
- W3035888673 hasConcept C33923547 @default.
- W3035888673 hasConcept C41008148 @default.
- W3035888673 hasConcept C49937458 @default.
- W3035888673 hasConcept C50644808 @default.
- W3035888673 hasConcept C68339613 @default.
- W3035888673 hasConcept C68781425 @default.
- W3035888673 hasConcept C81363708 @default.
- W3035888673 hasConcept C8642999 @default.
- W3035888673 hasConceptScore W3035888673C111919701 @default.
- W3035888673 hasConceptScore W3035888673C11413529 @default.
- W3035888673 hasConceptScore W3035888673C119857082 @default.
- W3035888673 hasConceptScore W3035888673C126255220 @default.
- W3035888673 hasConceptScore W3035888673C137635306 @default.
- W3035888673 hasConceptScore W3035888673C154945302 @default.
- W3035888673 hasConceptScore W3035888673C165464430 @default.
- W3035888673 hasConceptScore W3035888673C190502265 @default.
- W3035888673 hasConceptScore W3035888673C2776214188 @default.
- W3035888673 hasConceptScore W3035888673C2778049539 @default.
- W3035888673 hasConceptScore W3035888673C33923547 @default.
- W3035888673 hasConceptScore W3035888673C41008148 @default.
- W3035888673 hasConceptScore W3035888673C49937458 @default.
- W3035888673 hasConceptScore W3035888673C50644808 @default.
- W3035888673 hasConceptScore W3035888673C68339613 @default.
- W3035888673 hasConceptScore W3035888673C68781425 @default.
- W3035888673 hasConceptScore W3035888673C81363708 @default.
- W3035888673 hasConceptScore W3035888673C8642999 @default.
- W3035888673 hasLocation W30358886731 @default.
- W3035888673 hasOpenAccess W3035888673 @default.
- W3035888673 hasPrimaryLocation W30358886731 @default.
- W3035888673 hasRelatedWork W2949431720 @default.
- W3035888673 hasRelatedWork W2978290780 @default.
- W3035888673 hasRelatedWork W3192751261 @default.
- W3035888673 hasRelatedWork W3199608561 @default.
- W3035888673 hasRelatedWork W3216553692 @default.
- W3035888673 hasRelatedWork W4280606586 @default.
- W3035888673 hasRelatedWork W4287027148 @default.
- W3035888673 hasRelatedWork W4297803580 @default.
- W3035888673 hasRelatedWork W4311110689 @default.
- W3035888673 hasRelatedWork W4311110694 @default.
- W3035888673 isParatext "false" @default.
- W3035888673 isRetracted "false" @default.
- W3035888673 magId "3035888673" @default.
- W3035888673 workType "article" @default.