Matches in SemOpenAlex for { <https://semopenalex.org/work/W2782603654> ?p ?o ?g. }
- W2782603654 endingPage "145" @default.
- W2782603654 startingPage "131" @default.
- W2782603654 abstract "Relying on a single imperfect human annotator is not recommended in real crowdsourced classification problems. In practice, several annotators’ propositions are generally aggregated to obtain a better classification accuracy. Bayesian approaches, by modeling the relationship between each annotator’s output and the possible true labels (classes), have been shown to outperform other simpler models. Unfortunately, they assume that the total number of true labels is known. This is not the case in lots of realistic scenarios such as open-world classification where the number of possible labels is undetermined and may change over time. In this paper, we show how to set a non-parametric prior over the possible label set using the Dirichlet process in order to overcome this limitation. We illustrate this prior over the Bayesian annotator combination (BAC) model from the state of the art, resulting in the so-called non-parametric BAC (NPBAC). We show how to derive its variational equations to evaluate the model and how to assess it when the Dirichlet process has a prior using the Laplace method. We apply the model to several scenarios related to closed-world classification, open-world classification and novelty detection on a dataset previously published and on two datasets related to plant classification. Our experiments show that NPBAC is able to determine the true number of labels, but also and surprisingly, it largely outperforms the parametric annotator combination by modeling more complex confusions, in particular when few or no training data are available." @default.
- W2782603654 created "2018-01-26" @default.
- W2782603654 creator A5012815620 @default.
- W2782603654 creator A5015501462 @default.
- W2782603654 creator A5018620672 @default.
- W2782603654 date "2018-04-01" @default.
- W2782603654 modified "2023-09-27" @default.
- W2782603654 title "Non-parametric Bayesian annotator combination" @default.
- W2782603654 cites W1459599406 @default.
- W2782603654 cites W1542601440 @default.
- W2782603654 cites W1839751919 @default.
- W2782603654 cites W1920170702 @default.
- W2782603654 cites W1955368298 @default.
- W2782603654 cites W2010974761 @default.
- W2782603654 cites W2054716087 @default.
- W2782603654 cites W2069429561 @default.
- W2782603654 cites W2108598243 @default.
- W2782603654 cites W2109021302 @default.
- W2782603654 cites W2127498532 @default.
- W2782603654 cites W2134305421 @default.
- W2782603654 cites W2136634080 @default.
- W2782603654 cites W2140251433 @default.
- W2782603654 cites W2141649520 @default.
- W2782603654 cites W2141708418 @default.
- W2782603654 cites W2152009989 @default.
- W2782603654 cites W2157010107 @default.
- W2782603654 cites W2164124780 @default.
- W2782603654 cites W2295675274 @default.
- W2782603654 cites W2397631451 @default.
- W2782603654 cites W2538903535 @default.
- W2782603654 cites W2574793465 @default.
- W2782603654 cites W9014458 @default.
- W2782603654 cites W3145738572 @default.
- W2782603654 doi "https://doi.org/10.1016/j.ins.2018.01.020" @default.
- W2782603654 hasPublicationYear "2018" @default.
- W2782603654 type Work @default.
- W2782603654 sameAs 2782603654 @default.
- W2782603654 citedByCount "1" @default.
- W2782603654 countsByYear W27826036542020 @default.
- W2782603654 crossrefType "journal-article" @default.
- W2782603654 hasAuthorship W2782603654A5012815620 @default.
- W2782603654 hasAuthorship W2782603654A5015501462 @default.
- W2782603654 hasAuthorship W2782603654A5018620672 @default.
- W2782603654 hasBestOaLocation W27826036542 @default.
- W2782603654 hasConcept C105795698 @default.
- W2782603654 hasConcept C107673813 @default.
- W2782603654 hasConcept C111919701 @default.
- W2782603654 hasConcept C117251300 @default.
- W2782603654 hasConcept C119857082 @default.
- W2782603654 hasConcept C124101348 @default.
- W2782603654 hasConcept C134306372 @default.
- W2782603654 hasConcept C138885662 @default.
- W2782603654 hasConcept C153180895 @default.
- W2782603654 hasConcept C154945302 @default.
- W2782603654 hasConcept C169214877 @default.
- W2782603654 hasConcept C177264268 @default.
- W2782603654 hasConcept C182310444 @default.
- W2782603654 hasConcept C199360897 @default.
- W2782603654 hasConcept C27206212 @default.
- W2782603654 hasConcept C2778738651 @default.
- W2782603654 hasConcept C2778924833 @default.
- W2782603654 hasConcept C2780310539 @default.
- W2782603654 hasConcept C2781280628 @default.
- W2782603654 hasConcept C33923547 @default.
- W2782603654 hasConcept C41008148 @default.
- W2782603654 hasConcept C41895202 @default.
- W2782603654 hasConcept C98045186 @default.
- W2782603654 hasConceptScore W2782603654C105795698 @default.
- W2782603654 hasConceptScore W2782603654C107673813 @default.
- W2782603654 hasConceptScore W2782603654C111919701 @default.
- W2782603654 hasConceptScore W2782603654C117251300 @default.
- W2782603654 hasConceptScore W2782603654C119857082 @default.
- W2782603654 hasConceptScore W2782603654C124101348 @default.
- W2782603654 hasConceptScore W2782603654C134306372 @default.
- W2782603654 hasConceptScore W2782603654C138885662 @default.
- W2782603654 hasConceptScore W2782603654C153180895 @default.
- W2782603654 hasConceptScore W2782603654C154945302 @default.
- W2782603654 hasConceptScore W2782603654C169214877 @default.
- W2782603654 hasConceptScore W2782603654C177264268 @default.
- W2782603654 hasConceptScore W2782603654C182310444 @default.
- W2782603654 hasConceptScore W2782603654C199360897 @default.
- W2782603654 hasConceptScore W2782603654C27206212 @default.
- W2782603654 hasConceptScore W2782603654C2778738651 @default.
- W2782603654 hasConceptScore W2782603654C2778924833 @default.
- W2782603654 hasConceptScore W2782603654C2780310539 @default.
- W2782603654 hasConceptScore W2782603654C2781280628 @default.
- W2782603654 hasConceptScore W2782603654C33923547 @default.
- W2782603654 hasConceptScore W2782603654C41008148 @default.
- W2782603654 hasConceptScore W2782603654C41895202 @default.
- W2782603654 hasConceptScore W2782603654C98045186 @default.
- W2782603654 hasLocation W27826036541 @default.
- W2782603654 hasLocation W27826036542 @default.
- W2782603654 hasLocation W27826036543 @default.
- W2782603654 hasOpenAccess W2782603654 @default.
- W2782603654 hasPrimaryLocation W27826036541 @default.
- W2782603654 hasRelatedWork W1570269564 @default.
- W2782603654 hasRelatedWork W2027738917 @default.
- W2782603654 hasRelatedWork W2172000475 @default.