Matches in SemOpenAlex for { <https://semopenalex.org/work/W2945308135> ?p ?o ?g. }
- W2945308135 abstract "Designing an effective loss function plays an important role in visual analysis. Most existing loss function designs rely on hand-crafted heuristics that require domain experts to explore the large design space, which is usually sub-optimal and time-consuming. In this paper, we propose AutoML for Loss Function Search (AM-LFS) which leverages REINFORCE to search loss functions during the training process. The key contribution of this work is the design of search space which can guarantee the generalization and transferability on different vision tasks by including a bunch of existing prevailing loss functions in a unified formulation. We also propose an efficient optimization framework which can dynamically optimize the parameters of loss function's distribution during training. Extensive experimental results on four benchmark datasets show that, without any tricks, our method outperforms existing hand-crafted loss functions in various computer vision tasks." @default.
- W2945308135 created "2019-05-29" @default.
- W2945308135 creator A5018791285 @default.
- W2945308135 creator A5027254893 @default.
- W2945308135 creator A5054033403 @default.
- W2945308135 creator A5063854306 @default.
- W2945308135 creator A5072137495 @default.
- W2945308135 creator A5087073100 @default.
- W2945308135 creator A5087818121 @default.
- W2945308135 date "2019-05-17" @default.
- W2945308135 modified "2023-09-23" @default.
- W2945308135 title "AM-LFS: AutoML for Loss Function Search" @default.
- W2945308135 cites W1509966554 @default.
- W2945308135 cites W1536680647 @default.
- W2945308135 cites W1799366690 @default.
- W2945308135 cites W2119717200 @default.
- W2945308135 cites W2124659975 @default.
- W2945308135 cites W2132083787 @default.
- W2945308135 cites W2194775991 @default.
- W2945308135 cites W2204750386 @default.
- W2945308135 cites W2331143823 @default.
- W2945308135 cites W2511791013 @default.
- W2945308135 cites W2553303224 @default.
- W2945308135 cites W2585635281 @default.
- W2945308135 cites W2613718673 @default.
- W2945308135 cites W2782417188 @default.
- W2945308135 cites W2784874046 @default.
- W2945308135 cites W2785366763 @default.
- W2945308135 cites W2785430118 @default.
- W2945308135 cites W2786471719 @default.
- W2945308135 cites W2886851211 @default.
- W2945308135 cites W2888429796 @default.
- W2945308135 cites W2892359699 @default.
- W2945308135 cites W2901278409 @default.
- W2945308135 cites W2902251695 @default.
- W2945308135 cites W2905150051 @default.
- W2945308135 cites W2951104886 @default.
- W2945308135 cites W2962691289 @default.
- W2945308135 cites W2962706983 @default.
- W2945308135 cites W2962835968 @default.
- W2945308135 cites W2962898354 @default.
- W2945308135 cites W2962926870 @default.
- W2945308135 cites W2963322158 @default.
- W2945308135 cites W2963351448 @default.
- W2945308135 cites W2963466847 @default.
- W2945308135 cites W2963656735 @default.
- W2945308135 cites W2963671154 @default.
- W2945308135 cites W2963780286 @default.
- W2945308135 cites W2963842104 @default.
- W2945308135 cites W2963918968 @default.
- W2945308135 cites W2964044605 @default.
- W2945308135 cites W2964081403 @default.
- W2945308135 cites W2964081807 @default.
- W2945308135 cites W3100506510 @default.
- W2945308135 cites W3103152812 @default.
- W2945308135 cites W3118608800 @default.
- W2945308135 doi "https://doi.org/10.48550/arxiv.1905.07375" @default.
- W2945308135 hasPublicationYear "2019" @default.
- W2945308135 type Work @default.
- W2945308135 sameAs 2945308135 @default.
- W2945308135 citedByCount "6" @default.
- W2945308135 countsByYear W29453081352019 @default.
- W2945308135 countsByYear W29453081352020 @default.
- W2945308135 crossrefType "posted-content" @default.
- W2945308135 hasAuthorship W2945308135A5018791285 @default.
- W2945308135 hasAuthorship W2945308135A5027254893 @default.
- W2945308135 hasAuthorship W2945308135A5054033403 @default.
- W2945308135 hasAuthorship W2945308135A5063854306 @default.
- W2945308135 hasAuthorship W2945308135A5072137495 @default.
- W2945308135 hasAuthorship W2945308135A5087073100 @default.
- W2945308135 hasAuthorship W2945308135A5087818121 @default.
- W2945308135 hasBestOaLocation W29453081351 @default.
- W2945308135 hasConcept C111919701 @default.
- W2945308135 hasConcept C119857082 @default.
- W2945308135 hasConcept C127705205 @default.
- W2945308135 hasConcept C13280743 @default.
- W2945308135 hasConcept C134306372 @default.
- W2945308135 hasConcept C14036430 @default.
- W2945308135 hasConcept C154945302 @default.
- W2945308135 hasConcept C177148314 @default.
- W2945308135 hasConcept C185798385 @default.
- W2945308135 hasConcept C205649164 @default.
- W2945308135 hasConcept C26517878 @default.
- W2945308135 hasConcept C2778572836 @default.
- W2945308135 hasConcept C33923547 @default.
- W2945308135 hasConcept C36503486 @default.
- W2945308135 hasConcept C38652104 @default.
- W2945308135 hasConcept C41008148 @default.
- W2945308135 hasConcept C78458016 @default.
- W2945308135 hasConcept C86803240 @default.
- W2945308135 hasConcept C98045186 @default.
- W2945308135 hasConceptScore W2945308135C111919701 @default.
- W2945308135 hasConceptScore W2945308135C119857082 @default.
- W2945308135 hasConceptScore W2945308135C127705205 @default.
- W2945308135 hasConceptScore W2945308135C13280743 @default.
- W2945308135 hasConceptScore W2945308135C134306372 @default.
- W2945308135 hasConceptScore W2945308135C14036430 @default.
- W2945308135 hasConceptScore W2945308135C154945302 @default.
- W2945308135 hasConceptScore W2945308135C177148314 @default.
- W2945308135 hasConceptScore W2945308135C185798385 @default.