FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W2912168444> ?p ?o ?g. }

• W2912168444 abstract "We propose SWA-Gaussian (SWAG), a simple, scalable, and general purpose approach for uncertainty representation and calibration in deep learning. Stochastic Weight Averaging (SWA), which computes the first moment of stochastic gradient descent (SGD) iterates with a modified learning rate schedule, has recently been shown to improve generalization in deep learning. With SWAG, we fit a Gaussian using the SWA solution as the first moment and a low rank plus diagonal covariance also derived from the SGD iterates, forming an approximate posterior distribution over neural network weights; we then sample from this Gaussian distribution to perform Bayesian model averaging. We empirically find that SWAG approximates the shape of the true posterior, in accordance with results describing the stationary distribution of SGD iterates. Moreover, we demonstrate that SWAG performs well on a wide variety of tasks, including out of sample detection, calibration, and transfer learning, in comparison to many popular alternatives including MC dropout, KFAC Laplace, SGLD, and temperature scaling." @default.
• W2912168444 created "2019-02-21" @default.
• W2912168444 creator A5028940329 @default.
• W2912168444 creator A5032014309 @default.
• W2912168444 creator A5043360924 @default.
• W2912168444 creator A5066140003 @default.
• W2912168444 creator A5069234607 @default.
• W2912168444 date "2019-02-07" @default.
• W2912168444 modified "2023-09-27" @default.
• W2912168444 title "A Simple Baseline for Bayesian Uncertainty in Deep Learning" @default.
• W2912168444 cites W1492657231 @default.
• W2912168444 cites W1567512734 @default.
• W2912168444 cites W1826234144 @default.
• W2912168444 cites W1959608418 @default.
• W2912168444 cites W1988520084 @default.
• W2912168444 cites W2086161653 @default.
• W2912168444 cites W2098824882 @default.
• W2912168444 cites W2102685295 @default.
• W2912168444 cites W2108677974 @default.
• W2912168444 cites W2111051539 @default.
• W2912168444 cites W2117539524 @default.
• W2912168444 cites W2117756735 @default.
• W2912168444 cites W2118858186 @default.
• W2912168444 cites W2144193737 @default.
• W2912168444 cites W2164411961 @default.
• W2912168444 cites W2167433878 @default.
• W2912168444 cites W2194775991 @default.
• W2912168444 cites W2254249950 @default.
• W2912168444 cites W2257113116 @default.
• W2912168444 cites W2401231614 @default.
• W2912168444 cites W2545578326 @default.
• W2912168444 cites W2560647685 @default.
• W2912168444 cites W2592505114 @default.
• W2912168444 cites W2743945814 @default.
• W2912168444 cites W2752013927 @default.
• W2912168444 cites W2774412855 @default.
• W2912168444 cites W2786857698 @default.
• W2912168444 cites W2804017338 @default.
• W2912168444 cites W2893995718 @default.
• W2912168444 cites W2897001865 @default.
• W2912168444 cites W2903179019 @default.
• W2912168444 cites W2904243021 @default.
• W2912168444 cites W2908510526 @default.
• W2912168444 cites W2919841361 @default.
• W2912168444 cites W2940542208 @default.
• W2912168444 cites W2949117887 @default.
• W2912168444 cites W2950117601 @default.
• W2912168444 cites W2950177356 @default.
• W2912168444 cites W2950517871 @default.
• W2912168444 cites W2951196414 @default.
• W2912168444 cites W2951266961 @default.
• W2912168444 cites W2952088488 @default.
• W2912168444 cites W2953263857 @default.
• W2912168444 cites W2962712513 @default.
• W2912168444 cites W2962915600 @default.
• W2912168444 cites W2962933129 @default.
• W2912168444 cites W2963003887 @default.
• W2912168444 cites W2963008903 @default.
• W2912168444 cites W2963025848 @default.
• W2912168444 cites W2963027765 @default.
• W2912168444 cites W2963081736 @default.
• W2912168444 cites W2963173418 @default.
• W2912168444 cites W2963177640 @default.
• W2912168444 cites W2963190151 @default.
• W2912168444 cites W2963238274 @default.
• W2912168444 cites W2963384892 @default.
• W2912168444 cites W2963446712 @default.
• W2912168444 cites W2963783084 @default.
• W2912168444 cites W2963959597 @default.
• W2912168444 cites W2964212410 @default.
• W2912168444 cites W3037786240 @default.
• W2912168444 cites W3140968660 @default.
• W2912168444 cites W2959490256 @default.
• W2912168444 hasPublicationYear "2019" @default.
• W2912168444 type Work @default.
• W2912168444 sameAs 2912168444 @default.
• W2912168444 citedByCount "98" @default.
• W2912168444 countsByYear W29121684442019 @default.
• W2912168444 countsByYear W29121684442020 @default.
• W2912168444 countsByYear W29121684442021 @default.
• W2912168444 countsByYear W29121684442022 @default.
• W2912168444 crossrefType "posted-content" @default.
• W2912168444 hasAuthorship W2912168444A5028940329 @default.
• W2912168444 hasAuthorship W2912168444A5032014309 @default.
• W2912168444 hasAuthorship W2912168444A5043360924 @default.
• W2912168444 hasAuthorship W2912168444A5066140003 @default.
• W2912168444 hasAuthorship W2912168444A5069234607 @default.
• W2912168444 hasConcept C107673813 @default.
• W2912168444 hasConcept C11413529 @default.
• W2912168444 hasConcept C121332964 @default.
• W2912168444 hasConcept C126255220 @default.
• W2912168444 hasConcept C134306372 @default.
• W2912168444 hasConcept C140479938 @default.
• W2912168444 hasConcept C154945302 @default.
• W2912168444 hasConcept C163716315 @default.
• W2912168444 hasConcept C206688291 @default.
• W2912168444 hasConcept C22243797 @default.
• W2912168444 hasConcept C28826006 @default.
• W2912168444 hasConcept C33923547 @default.
• W2912168444 hasConcept C41008148 @default.

• next