Matches in SemOpenAlex for { <https://semopenalex.org/work/W4381587777> ?p ?o ?g. }
Showing items 1 to 81 of
81
with 100 items per page.
- W4381587777 abstract "Unsupervised performance estimation, or evaluating how well models perform on unlabeled data is a difficult task. Recently, a method was proposed by Garg et al. [2022] which performs much better than previous methods. Their method relies on having a score function, satisfying certain properties, to map probability vectors outputted by the classifier to the reals, but it is an open problem which score function is best. We explore this problem by first showing that their method fundamentally relies on the ordering induced by this score function. Thus, under monotone transformations of score functions, their method yields the same estimate. Next, we show that in the binary classification setting, nearly all common score functions - the $L^infty$ norm; the $L^2$ norm; negative entropy; and the $L^2$, $L^1$, and Jensen-Shannon distances to the uniform vector - all induce the same ordering over probability vectors. However, this does not hold for higher dimensional settings. We conduct numerous experiments on well-known NLP data sets and rigorously explore the performance of different score functions. We conclude that the $L^infty$ norm is the most appropriate." @default.
- W4381587777 created "2023-06-22" @default.
- W4381587777 creator A5025617226 @default.
- W4381587777 creator A5034370385 @default.
- W4381587777 creator A5061336404 @default.
- W4381587777 creator A5088612057 @default.
- W4381587777 date "2023-06-16" @default.
- W4381587777 modified "2023-10-18" @default.
- W4381587777 title "On Orderings of Probability Vectors and Unsupervised Performance Estimation" @default.
- W4381587777 doi "https://doi.org/10.48550/arxiv.2306.10160" @default.
- W4381587777 hasPublicationYear "2023" @default.
- W4381587777 type Work @default.
- W4381587777 citedByCount "0" @default.
- W4381587777 crossrefType "posted-content" @default.
- W4381587777 hasAuthorship W4381587777A5025617226 @default.
- W4381587777 hasAuthorship W4381587777A5034370385 @default.
- W4381587777 hasAuthorship W4381587777A5061336404 @default.
- W4381587777 hasAuthorship W4381587777A5088612057 @default.
- W4381587777 hasBestOaLocation W43815877771 @default.
- W4381587777 hasConcept C105795698 @default.
- W4381587777 hasConcept C106301342 @default.
- W4381587777 hasConcept C11413529 @default.
- W4381587777 hasConcept C121332964 @default.
- W4381587777 hasConcept C12267149 @default.
- W4381587777 hasConcept C14036430 @default.
- W4381587777 hasConcept C153180895 @default.
- W4381587777 hasConcept C154945302 @default.
- W4381587777 hasConcept C17744445 @default.
- W4381587777 hasConcept C191795146 @default.
- W4381587777 hasConcept C199539241 @default.
- W4381587777 hasConcept C2524010 @default.
- W4381587777 hasConcept C2834757 @default.
- W4381587777 hasConcept C28826006 @default.
- W4381587777 hasConcept C33923547 @default.
- W4381587777 hasConcept C41008148 @default.
- W4381587777 hasConcept C48372109 @default.
- W4381587777 hasConcept C62520636 @default.
- W4381587777 hasConcept C65660741 @default.
- W4381587777 hasConcept C66905080 @default.
- W4381587777 hasConcept C78458016 @default.
- W4381587777 hasConcept C86803240 @default.
- W4381587777 hasConcept C94375191 @default.
- W4381587777 hasConceptScore W4381587777C105795698 @default.
- W4381587777 hasConceptScore W4381587777C106301342 @default.
- W4381587777 hasConceptScore W4381587777C11413529 @default.
- W4381587777 hasConceptScore W4381587777C121332964 @default.
- W4381587777 hasConceptScore W4381587777C12267149 @default.
- W4381587777 hasConceptScore W4381587777C14036430 @default.
- W4381587777 hasConceptScore W4381587777C153180895 @default.
- W4381587777 hasConceptScore W4381587777C154945302 @default.
- W4381587777 hasConceptScore W4381587777C17744445 @default.
- W4381587777 hasConceptScore W4381587777C191795146 @default.
- W4381587777 hasConceptScore W4381587777C199539241 @default.
- W4381587777 hasConceptScore W4381587777C2524010 @default.
- W4381587777 hasConceptScore W4381587777C2834757 @default.
- W4381587777 hasConceptScore W4381587777C28826006 @default.
- W4381587777 hasConceptScore W4381587777C33923547 @default.
- W4381587777 hasConceptScore W4381587777C41008148 @default.
- W4381587777 hasConceptScore W4381587777C48372109 @default.
- W4381587777 hasConceptScore W4381587777C62520636 @default.
- W4381587777 hasConceptScore W4381587777C65660741 @default.
- W4381587777 hasConceptScore W4381587777C66905080 @default.
- W4381587777 hasConceptScore W4381587777C78458016 @default.
- W4381587777 hasConceptScore W4381587777C86803240 @default.
- W4381587777 hasConceptScore W4381587777C94375191 @default.
- W4381587777 hasLocation W43815877771 @default.
- W4381587777 hasOpenAccess W4381587777 @default.
- W4381587777 hasPrimaryLocation W43815877771 @default.
- W4381587777 hasRelatedWork W2041399278 @default.
- W4381587777 hasRelatedWork W2099369243 @default.
- W4381587777 hasRelatedWork W2136184105 @default.
- W4381587777 hasRelatedWork W2141705618 @default.
- W4381587777 hasRelatedWork W2348964713 @default.
- W4381587777 hasRelatedWork W2970229296 @default.
- W4381587777 hasRelatedWork W4223656335 @default.
- W4381587777 hasRelatedWork W4285281467 @default.
- W4381587777 hasRelatedWork W2187500075 @default.
- W4381587777 hasRelatedWork W2345184372 @default.
- W4381587777 isParatext "false" @default.
- W4381587777 isRetracted "false" @default.
- W4381587777 workType "article" @default.