FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W2960055608> ?p ?o ?g. }

• W2960055608 abstract "Learning unbiased models on imbalanced datasets is a significant challenge. Rare classes tend to get a concentrated representation in the classification space which hampers the generalization of learned boundaries to new test examples. In this paper, we demonstrate that the Bayesian uncertainty estimates directly correlate with the rarity of classes and the difficulty level of individual samples. Subsequently, we present a novel framework for uncertainty based class imbalance learning that follows two key insights: First, classification boundaries should be extended further away from a more uncertain (rare) class to avoid overfitting and enhance its generalization. Second, each sample should be modeled as a multi-variate Gaussian distribution with a mean vector and a covariance matrix defined by the sample's uncertainty. The learned boundaries should respect not only the individual samples but also their distribution in the feature space. Our proposed approach efficiently utilizes sample and class uncertainty information to learn robust features and more generalizable classifiers. We systematically study the class imbalance problem and derive a novel loss formulation for max-margin learning based on Bayesian uncertainty measure. The proposed method shows significant performance improvements on six benchmark datasets for face verification, attribute prediction, digit/object classification and skin lesion detection." @default.
• W2960055608 created "2019-07-23" @default.
• W2960055608 creator A5000300751 @default.
• W2960055608 creator A5041437150 @default.
• W2960055608 creator A5070709035 @default.
• W2960055608 creator A5074536754 @default.
• W2960055608 creator A5082634513 @default.
• W2960055608 date "2019-01-22" @default.
• W2960055608 modified "2023-10-15" @default.
• W2960055608 title "Striking the Right Balance with Uncertainty" @default.
• W2960055608 cites W1480009832 @default.
• W2960055608 cites W1502922572 @default.
• W2960055608 cites W1551909886 @default.
• W2960055608 cites W1581587400 @default.
• W2960055608 cites W1703179648 @default.
• W2960055608 cites W1834627138 @default.
• W2960055608 cites W1950843348 @default.
• W2960055608 cites W1994961181 @default.
• W2960055608 cites W2019464758 @default.
• W2960055608 cites W2022477494 @default.
• W2960055608 cites W2095705004 @default.
• W2960055608 cites W2104933073 @default.
• W2960055608 cites W2107138773 @default.
• W2960055608 cites W2108598243 @default.
• W2960055608 cites W2118978333 @default.
• W2960055608 cites W2128965734 @default.
• W2960055608 cites W2131427446 @default.
• W2960055608 cites W2136903812 @default.
• W2960055608 cites W2137029138 @default.
• W2960055608 cites W2144172034 @default.
• W2960055608 cites W2145287260 @default.
• W2960055608 cites W2147414309 @default.
• W2960055608 cites W2148143831 @default.
• W2960055608 cites W2157720605 @default.
• W2960055608 cites W2243526054 @default.
• W2960055608 cites W2253535400 @default.
• W2960055608 cites W2404498690 @default.
• W2960055608 cites W2440599146 @default.
• W2960055608 cites W2520774990 @default.
• W2960055608 cites W2594867206 @default.
• W2960055608 cites W2609296554 @default.
• W2960055608 cites W2618011341 @default.
• W2960055608 cites W2737608545 @default.
• W2960055608 cites W2752782242 @default.
• W2960055608 cites W2766890262 @default.
• W2960055608 cites W2781292787 @default.
• W2960055608 cites W2784874046 @default.
• W2960055608 cites W2791890924 @default.
• W2960055608 cites W2794467005 @default.
• W2960055608 cites W2795061970 @default.
• W2960055608 cites W2911606395 @default.
• W2960055608 cites W2951552869 @default.
• W2960055608 cites W2962835968 @default.
• W2960055608 cites W2962898354 @default.
• W2960055608 cites W2963227127 @default.
• W2960055608 cites W2963371670 @default.
• W2960055608 cites W2963656735 @default.
• W2960055608 cites W2963677766 @default.
• W2960055608 cites W2963839617 @default.
• W2960055608 cites W2963919294 @default.
• W2960055608 cites W2964050365 @default.
• W2960055608 cites W2964059111 @default.
• W2960055608 cites W2967324748 @default.
• W2960055608 cites W2971487614 @default.
• W2960055608 cites W3101998545 @default.
• W2960055608 cites W3103152812 @default.
• W2960055608 cites W3155649056 @default.
• W2960055608 cites W728297 @default.
• W2960055608 doi "https://doi.org/10.48550/arxiv.1901.07590" @default.
• W2960055608 hasPublicationYear "2019" @default.
• W2960055608 type Work @default.
• W2960055608 sameAs 2960055608 @default.
• W2960055608 citedByCount "17" @default.
• W2960055608 countsByYear W29600556082019 @default.
• W2960055608 countsByYear W29600556082020 @default.
• W2960055608 countsByYear W29600556082021 @default.
• W2960055608 crossrefType "posted-content" @default.
• W2960055608 hasAuthorship W2960055608A5000300751 @default.
• W2960055608 hasAuthorship W2960055608A5041437150 @default.
• W2960055608 hasAuthorship W2960055608A5070709035 @default.
• W2960055608 hasAuthorship W2960055608A5074536754 @default.
• W2960055608 hasAuthorship W2960055608A5082634513 @default.
• W2960055608 hasBestOaLocation W29600556081 @default.
• W2960055608 hasConcept C107673813 @default.
• W2960055608 hasConcept C119857082 @default.
• W2960055608 hasConcept C13280743 @default.
• W2960055608 hasConcept C134306372 @default.
• W2960055608 hasConcept C138885662 @default.
• W2960055608 hasConcept C153180895 @default.
• W2960055608 hasConcept C154945302 @default.
• W2960055608 hasConcept C177148314 @default.
• W2960055608 hasConcept C17744445 @default.
• W2960055608 hasConcept C185592680 @default.
• W2960055608 hasConcept C185798385 @default.
• W2960055608 hasConcept C198531522 @default.
• W2960055608 hasConcept C199539241 @default.
• W2960055608 hasConcept C205649164 @default.
• W2960055608 hasConcept C22019652 @default.
• W2960055608 hasConcept C2776359362 @default.
• W2960055608 hasConcept C2776401178 @default.

• next