Matches in SemOpenAlex for { <https://semopenalex.org/work/W2725281132> ?p ?o ?g. }
- W2725281132 abstract "Computer vision is experiencing an AI renaissance, in which machine learning models are expediting important breakthroughs in academic research and commercial applications. Effectively training these models, however, is not trivial due in part to hyperparameters: user-configured values that control a model's ability to learn from data. Existing hyperparameter optimization methods are highly parallel but make no effort to balance the search across heterogeneous hardware or to prioritize searching high-impact spaces. In this paper, we introduce a framework for massively Scalable Hardware-Aware Distributed Hyperparameter Optimization (SHADHO). Our framework calculates the relative complexity of each search space and monitors performance on the learning task over all trials. These metrics are then used as heuristics to assign hyperparameters to distributed workers based on their hardware. We first demonstrate that our framework achieves double the throughput of a standard distributed hyperparameter optimization framework by optimizing SVM for MNIST using 150 distributed workers. We then conduct model search with SHADHO over the course of one week using 74 GPUs across two compute clusters to optimize U-Net for a cell segmentation task, discovering 515 models that achieve a lower validation loss than standard U-Net." @default.
- W2725281132 created "2017-07-14" @default.
- W2725281132 creator A5007301255 @default.
- W2725281132 creator A5054916731 @default.
- W2725281132 creator A5079809848 @default.
- W2725281132 creator A5091513740 @default.
- W2725281132 date "2017-07-05" @default.
- W2725281132 modified "2023-09-27" @default.
- W2725281132 title "SHADHO: Massively Scalable Hardware-Aware Distributed Hyperparameter Optimization" @default.
- W2725281132 cites W1437335841 @default.
- W2725281132 cites W1510052597 @default.
- W2725281132 cites W1513082520 @default.
- W2725281132 cites W1524673385 @default.
- W2725281132 cites W1901129140 @default.
- W2725281132 cites W191027614 @default.
- W2725281132 cites W1981713285 @default.
- W2725281132 cites W1986490585 @default.
- W2725281132 cites W1994197834 @default.
- W2725281132 cites W2016323639 @default.
- W2725281132 cites W2026354831 @default.
- W2725281132 cites W2097998348 @default.
- W2725281132 cites W2098368939 @default.
- W2725281132 cites W2101234009 @default.
- W2725281132 cites W2106411961 @default.
- W2725281132 cites W2110008785 @default.
- W2725281132 cites W2115755783 @default.
- W2725281132 cites W2119733057 @default.
- W2725281132 cites W2119821739 @default.
- W2725281132 cites W2143031615 @default.
- W2725281132 cites W2143192733 @default.
- W2725281132 cites W2149294210 @default.
- W2725281132 cites W2168338507 @default.
- W2725281132 cites W2172073485 @default.
- W2725281132 cites W2189149359 @default.
- W2725281132 cites W2266822037 @default.
- W2725281132 cites W2325939864 @default.
- W2725281132 cites W2342840547 @default.
- W2725281132 cites W2400200933 @default.
- W2725281132 cites W2464772092 @default.
- W2725281132 cites W2495425901 @default.
- W2725281132 cites W2550668432 @default.
- W2725281132 cites W2553303224 @default.
- W2725281132 cites W2581082771 @default.
- W2725281132 cites W2608595939 @default.
- W2725281132 cites W2660889552 @default.
- W2725281132 cites W2748513770 @default.
- W2725281132 cites W2750150650 @default.
- W2725281132 cites W2752790924 @default.
- W2725281132 cites W2794284562 @default.
- W2725281132 cites W2950182411 @default.
- W2725281132 cites W2964081807 @default.
- W2725281132 cites W2964234718 @default.
- W2725281132 hasPublicationYear "2017" @default.
- W2725281132 type Work @default.
- W2725281132 sameAs 2725281132 @default.
- W2725281132 citedByCount "0" @default.
- W2725281132 crossrefType "posted-content" @default.
- W2725281132 hasAuthorship W2725281132A5007301255 @default.
- W2725281132 hasAuthorship W2725281132A5054916731 @default.
- W2725281132 hasAuthorship W2725281132A5079809848 @default.
- W2725281132 hasAuthorship W2725281132A5091513740 @default.
- W2725281132 hasConcept C10485038 @default.
- W2725281132 hasConcept C108583219 @default.
- W2725281132 hasConcept C111919701 @default.
- W2725281132 hasConcept C119857082 @default.
- W2725281132 hasConcept C12267149 @default.
- W2725281132 hasConcept C127413603 @default.
- W2725281132 hasConcept C127705205 @default.
- W2725281132 hasConcept C134448949 @default.
- W2725281132 hasConcept C154945302 @default.
- W2725281132 hasConcept C157764524 @default.
- W2725281132 hasConcept C162324750 @default.
- W2725281132 hasConcept C173608175 @default.
- W2725281132 hasConcept C187736073 @default.
- W2725281132 hasConcept C190475519 @default.
- W2725281132 hasConcept C190502265 @default.
- W2725281132 hasConcept C201995342 @default.
- W2725281132 hasConcept C2780451532 @default.
- W2725281132 hasConcept C41008148 @default.
- W2725281132 hasConcept C48044578 @default.
- W2725281132 hasConcept C555944384 @default.
- W2725281132 hasConcept C76155785 @default.
- W2725281132 hasConcept C77088390 @default.
- W2725281132 hasConcept C8642999 @default.
- W2725281132 hasConceptScore W2725281132C10485038 @default.
- W2725281132 hasConceptScore W2725281132C108583219 @default.
- W2725281132 hasConceptScore W2725281132C111919701 @default.
- W2725281132 hasConceptScore W2725281132C119857082 @default.
- W2725281132 hasConceptScore W2725281132C12267149 @default.
- W2725281132 hasConceptScore W2725281132C127413603 @default.
- W2725281132 hasConceptScore W2725281132C127705205 @default.
- W2725281132 hasConceptScore W2725281132C134448949 @default.
- W2725281132 hasConceptScore W2725281132C154945302 @default.
- W2725281132 hasConceptScore W2725281132C157764524 @default.
- W2725281132 hasConceptScore W2725281132C162324750 @default.
- W2725281132 hasConceptScore W2725281132C173608175 @default.
- W2725281132 hasConceptScore W2725281132C187736073 @default.
- W2725281132 hasConceptScore W2725281132C190475519 @default.
- W2725281132 hasConceptScore W2725281132C190502265 @default.
- W2725281132 hasConceptScore W2725281132C201995342 @default.