Matches in SemOpenAlex for { <https://semopenalex.org/work/W2609862872> ?p ?o ?g. }
Showing items 1 to 78 of
78
with 100 items per page.
- W2609862872 abstract "Retrieving a small set of relevant and interesting objects from a large background class is challenging because classifiers can easily be overwhelmed by the large class. Classifiers have been developed that are more sensitive to the small class, and typically they optimize a ranking, or precision at the top. These measures can be costly because they often look at pairwise rankings. The classical approach of just reweighing the relevant objects also has its limits because the influence of outliers and mislabeled objects also dramatically increases, deteriorating the performance. In this paper we propose an alternative solution that uses non-convex and class dependent loss functions. The non-convex loss makes the classifier less sensitive to outliers, while the class-dependent loss stresses the interesting class. It can not only be used to solve retrieval problems, but also classification problems in which not all objects are labeled reliably, like in Multiple Instance Learning or Positive and Unlabeled data learning. For Multiple Instance Learning and learning from Positive and Unlabeled data it is even shown that these non-convex, class-dependent losses are already implicitly used." @default.
- W2609862872 created "2017-05-05" @default.
- W2609862872 creator A5001295629 @default.
- W2609862872 creator A5040922148 @default.
- W2609862872 date "2016-12-01" @default.
- W2609862872 modified "2023-09-24" @default.
- W2609862872 title "Class-dependent, non-convex losses to optimize precision" @default.
- W2609862872 cites W1556531089 @default.
- W2609862872 cites W1825821140 @default.
- W2609862872 cites W2107890099 @default.
- W2609862872 cites W2110119381 @default.
- W2609862872 cites W2123958887 @default.
- W2609862872 cites W2136688338 @default.
- W2609862872 cites W2137917285 @default.
- W2609862872 cites W2154318594 @default.
- W2609862872 doi "https://doi.org/10.1109/icpr.2016.7900145" @default.
- W2609862872 hasPublicationYear "2016" @default.
- W2609862872 type Work @default.
- W2609862872 sameAs 2609862872 @default.
- W2609862872 citedByCount "0" @default.
- W2609862872 crossrefType "proceedings-article" @default.
- W2609862872 hasAuthorship W2609862872A5001295629 @default.
- W2609862872 hasAuthorship W2609862872A5040922148 @default.
- W2609862872 hasConcept C112680207 @default.
- W2609862872 hasConcept C119857082 @default.
- W2609862872 hasConcept C153180895 @default.
- W2609862872 hasConcept C154945302 @default.
- W2609862872 hasConcept C184898388 @default.
- W2609862872 hasConcept C189430467 @default.
- W2609862872 hasConcept C2524010 @default.
- W2609862872 hasConcept C2777212361 @default.
- W2609862872 hasConcept C33923547 @default.
- W2609862872 hasConcept C34872919 @default.
- W2609862872 hasConcept C41008148 @default.
- W2609862872 hasConcept C79337645 @default.
- W2609862872 hasConcept C86037889 @default.
- W2609862872 hasConcept C95623464 @default.
- W2609862872 hasConceptScore W2609862872C112680207 @default.
- W2609862872 hasConceptScore W2609862872C119857082 @default.
- W2609862872 hasConceptScore W2609862872C153180895 @default.
- W2609862872 hasConceptScore W2609862872C154945302 @default.
- W2609862872 hasConceptScore W2609862872C184898388 @default.
- W2609862872 hasConceptScore W2609862872C189430467 @default.
- W2609862872 hasConceptScore W2609862872C2524010 @default.
- W2609862872 hasConceptScore W2609862872C2777212361 @default.
- W2609862872 hasConceptScore W2609862872C33923547 @default.
- W2609862872 hasConceptScore W2609862872C34872919 @default.
- W2609862872 hasConceptScore W2609862872C41008148 @default.
- W2609862872 hasConceptScore W2609862872C79337645 @default.
- W2609862872 hasConceptScore W2609862872C86037889 @default.
- W2609862872 hasConceptScore W2609862872C95623464 @default.
- W2609862872 hasLocation W26098628721 @default.
- W2609862872 hasOpenAccess W2609862872 @default.
- W2609862872 hasPrimaryLocation W26098628721 @default.
- W2609862872 hasRelatedWork W1525219522 @default.
- W2609862872 hasRelatedWork W1553533669 @default.
- W2609862872 hasRelatedWork W156764137 @default.
- W2609862872 hasRelatedWork W1753659477 @default.
- W2609862872 hasRelatedWork W1806550384 @default.
- W2609862872 hasRelatedWork W1948896869 @default.
- W2609862872 hasRelatedWork W1971595496 @default.
- W2609862872 hasRelatedWork W1977640460 @default.
- W2609862872 hasRelatedWork W2024847174 @default.
- W2609862872 hasRelatedWork W2040775130 @default.
- W2609862872 hasRelatedWork W2070101775 @default.
- W2609862872 hasRelatedWork W2113741880 @default.
- W2609862872 hasRelatedWork W2117345453 @default.
- W2609862872 hasRelatedWork W2243049275 @default.
- W2609862872 hasRelatedWork W2419570712 @default.
- W2609862872 hasRelatedWork W2495778891 @default.
- W2609862872 hasRelatedWork W2511672271 @default.
- W2609862872 hasRelatedWork W2804148081 @default.
- W2609862872 hasRelatedWork W3199262903 @default.
- W2609862872 hasRelatedWork W3089406873 @default.
- W2609862872 isParatext "false" @default.
- W2609862872 isRetracted "false" @default.
- W2609862872 magId "2609862872" @default.
- W2609862872 workType "article" @default.