Publication:
Simple, Robust, Scalable Semi-supervised Learning via Expectation Regularization

Simple item page
dc.contributor.authorMann, Gideon S.
dc.contributor.departmentUniversity of Massachusetts Amherst
dc.date2023-09-22T21:10:59.000
dc.date.accessioned2024-04-26T09:33:04Z
dc.date.available2024-04-26T09:33:04Z
dc.date.issued2007-01-01
dc.descriptionThis paper was harvested from CiteSeer
dc.description.abstractAlthough semi-supervised learning has been an active area of research, its use in deployed applications is still relatively rare because the methods are often difficult to implement, fragile in tuning, or lacking in scalability. This paper presents expectation regularization, a semi-supervised learning method for exponential family parametric models that augments the traditional conditional label-likelihood objective function with an additional term that encourages model predictions on unlabeled data to match certain expectations—such as label priors. The method is extremely easy to implement, scales as well as logistic regression, and can handle non-independent features. We present experiments on five different data sets, showing accuracy improvements over other semi-supervised methods.
dc.identifier.urihttps://hdl.handle.net/20.500.14394/9532
dc.relation.urlhttps://scholarworks.umass.edu/cgi/viewcontent.cgi?article=1102&context=cs_faculty_pubs&unstamped=1
dc.source.statuspublished
dc.subjectComputer Sciences
dc.titleSimple, Robust, Scalable Semi-supervised Learning via Expectation Regularization
dc.typearticle
dc.typearticle
digcom.contributor.authorMann, Gideon S.
digcom.identifiercs_faculty_pubs/103
digcom.identifier.contextkey1300993
digcom.identifier.submissionpathcs_faculty_pubs/103
dspace.entity.typePublication
Files
Original bundle
Now showing 1 - 1 of 1
Thumbnail Image
Name:
Gideon_S_Mann_2.pdf
Size:
349.21 KB
Format:
Adobe Portable Document Format
Thumbnail Image
Name: Gideon_S_Mann_2.pdf
Size: 349.21 KB
Kind: Adobe PDF
Collections
Computer Science Department Faculty Publication Series