Show simple item record

Transfer learning for image classification with sparse prototype representations

dc.date.accessioned2008-03-03T14:45:13Z
dc.date.accessioned2018-11-26T22:25:11Z
dc.date.available2008-03-03T14:45:13Z
dc.date.available2018-11-26T22:25:11Z
dc.date.issued2008-03-03en_US
dc.identifier.urihttp://hdl.handle.net/1721.1/40797
dc.identifier.urihttp://repository.aust.edu.ng/xmlui/handle/1721.1/40797
dc.description.abstractTo learn a new visual category from few examples, prior knowledge from unlabeled data as well as previous related categories may be useful.  We develop a new method for transfer learning which exploits available unlabeled data and an arbitrary kernel function; we form a representation based on kernel distances to a large set of unlabeled data points. To transfer knowledge from previous related problems we observe that a category might be learnable using only a small subset of reference prototypes. Related problems may share a significant number of relevant prototypes; we find such a reduced representation by performing a joint loss minimization over the training sets of related problems with a shared regularization penalty that minimizes the total number of prototypes involved in the approximation.This optimization problem can be formulated as a linear program thatcan be solved efficiently. We conduct experiments on a news-topic prediction task where the goal is to predict whether an image belongs to a particularnews topic. Our results show that when only few examples are available for training a target topic, leveraging knowledge learnt from other topics can significantly improve performance.en_US
dc.format.extent8 p.en_US
dc.relationMassachusetts Institute of Technology Computer Science and Artificial Intelligence Laboratoryen_US
dc.relationen_US
dc.subjecttransfer learningen_US
dc.subjectimage classificationen_US
dc.titleTransfer learning for image classification with sparse prototype representationsen_US


Files in this item

FilesSizeFormatView
MIT-CSAIL-TR-2008-012.pdf239.7Kbapplication/pdfView/Open
MIT-CSAIL-TR-2008-012.ps4.693Mbapplication/postscriptView/Open

This item appears in the following Collection(s)

Show simple item record