Matches in SemOpenAlex for { <https://semopenalex.org/work/W4385825379> ?p ?o ?g. }
Showing items 1 to 79 of
79
with 100 items per page.
- W4385825379 abstract "Traditional deep learning (DL) models are powerful classifiers, but many approaches do not provide uncertainties for their estimates. Uncertainty quantification (UQ) methods for DL models have received increased attention in the literature due to their usefulness in decision making, particularly for high-consequence decisions. However, there has been little research done on how to evaluate the quality of such methods. We use statistical methods of frequentist interval coverage and interval width to evaluate the quality of credible intervals, and expected calibration error to evaluate classification predicted confidence. These metrics are evaluated on Bayesian neural networks (BNN) fit using Markov Chain Monte Carlo (MCMC) and variational inference (VI), bootstrapped neural networks (NN), Deep Ensembles (DE), and Monte Carlo (MC) dropout. We apply these different UQ for DL methods to a hyperspectral image target detection problem and show the inconsistency of the different methods' results and the necessity of a UQ quality metric. To reconcile these differences and choose a UQ method that appropriately quantifies the uncertainty, we create a simulated data set with fully parameterized probability distribution for a two-class classification problem. The gold standard MCMC performs the best overall, and the bootstrapped NN is a close second, requiring the same computational expense as DE. Through this comparison, we demonstrate that, for a given data set, different models can produce uncertainty estimates of markedly different quality. This in turn points to a great need for principled assessment methods of UQ quality in DL applications." @default.
- W4385825379 created "2023-08-15" @default.
- W4385825379 creator A5023459618 @default.
- W4385825379 creator A5024908504 @default.
- W4385825379 creator A5052665440 @default.
- W4385825379 creator A5075985334 @default.
- W4385825379 creator A5084163286 @default.
- W4385825379 date "2023-08-10" @default.
- W4385825379 modified "2023-09-27" @default.
- W4385825379 title "Comparing the quality of neural network uncertainty estimates for classification problems" @default.
- W4385825379 doi "https://doi.org/10.48550/arxiv.2308.05903" @default.
- W4385825379 hasPublicationYear "2023" @default.
- W4385825379 type Work @default.
- W4385825379 citedByCount "0" @default.
- W4385825379 crossrefType "posted-content" @default.
- W4385825379 hasAuthorship W4385825379A5023459618 @default.
- W4385825379 hasAuthorship W4385825379A5024908504 @default.
- W4385825379 hasAuthorship W4385825379A5052665440 @default.
- W4385825379 hasAuthorship W4385825379A5075985334 @default.
- W4385825379 hasAuthorship W4385825379A5084163286 @default.
- W4385825379 hasBestOaLocation W43858253791 @default.
- W4385825379 hasConcept C105795698 @default.
- W4385825379 hasConcept C107673813 @default.
- W4385825379 hasConcept C111350023 @default.
- W4385825379 hasConcept C11413529 @default.
- W4385825379 hasConcept C119857082 @default.
- W4385825379 hasConcept C124101348 @default.
- W4385825379 hasConcept C154945302 @default.
- W4385825379 hasConcept C160234255 @default.
- W4385825379 hasConcept C162324750 @default.
- W4385825379 hasConcept C162376815 @default.
- W4385825379 hasConcept C165838908 @default.
- W4385825379 hasConcept C176217482 @default.
- W4385825379 hasConcept C177264268 @default.
- W4385825379 hasConcept C19499675 @default.
- W4385825379 hasConcept C199360897 @default.
- W4385825379 hasConcept C21547014 @default.
- W4385825379 hasConcept C2776214188 @default.
- W4385825379 hasConcept C32230216 @default.
- W4385825379 hasConcept C33923547 @default.
- W4385825379 hasConcept C41008148 @default.
- W4385825379 hasConcept C50644808 @default.
- W4385825379 hasConceptScore W4385825379C105795698 @default.
- W4385825379 hasConceptScore W4385825379C107673813 @default.
- W4385825379 hasConceptScore W4385825379C111350023 @default.
- W4385825379 hasConceptScore W4385825379C11413529 @default.
- W4385825379 hasConceptScore W4385825379C119857082 @default.
- W4385825379 hasConceptScore W4385825379C124101348 @default.
- W4385825379 hasConceptScore W4385825379C154945302 @default.
- W4385825379 hasConceptScore W4385825379C160234255 @default.
- W4385825379 hasConceptScore W4385825379C162324750 @default.
- W4385825379 hasConceptScore W4385825379C162376815 @default.
- W4385825379 hasConceptScore W4385825379C165838908 @default.
- W4385825379 hasConceptScore W4385825379C176217482 @default.
- W4385825379 hasConceptScore W4385825379C177264268 @default.
- W4385825379 hasConceptScore W4385825379C19499675 @default.
- W4385825379 hasConceptScore W4385825379C199360897 @default.
- W4385825379 hasConceptScore W4385825379C21547014 @default.
- W4385825379 hasConceptScore W4385825379C2776214188 @default.
- W4385825379 hasConceptScore W4385825379C32230216 @default.
- W4385825379 hasConceptScore W4385825379C33923547 @default.
- W4385825379 hasConceptScore W4385825379C41008148 @default.
- W4385825379 hasConceptScore W4385825379C50644808 @default.
- W4385825379 hasLocation W43858253791 @default.
- W4385825379 hasOpenAccess W4385825379 @default.
- W4385825379 hasPrimaryLocation W43858253791 @default.
- W4385825379 hasRelatedWork W1860392020 @default.
- W4385825379 hasRelatedWork W1971450106 @default.
- W4385825379 hasRelatedWork W2951176680 @default.
- W4385825379 hasRelatedWork W3002319139 @default.
- W4385825379 hasRelatedWork W3134187673 @default.
- W4385825379 hasRelatedWork W3161293650 @default.
- W4385825379 hasRelatedWork W3167474566 @default.
- W4385825379 hasRelatedWork W4300815303 @default.
- W4385825379 hasRelatedWork W4364383560 @default.
- W4385825379 hasRelatedWork W4286903091 @default.
- W4385825379 isParatext "false" @default.
- W4385825379 isRetracted "false" @default.
- W4385825379 workType "article" @default.