Sidestepping Intractable Inference with Structured Ensemble Cascades

Simple item page
dc.contributor.authorWeiss, David
dc.contributor.authorSapp, Benjamin
dc.contributor.authorTaskar, Ben
dc.date2023-05-17T07:08:16.000
dc.date.accessioned2023-05-22T12:49:03Z
dc.date.available2023-05-22T12:49:03Z
dc.date.issued2010-12-01
dc.date.submitted2012-07-03T09:18:53-07:00
dc.description.abstractFor many structured prediction problems, complex models often require adopting approximate inference techniques such as variational methods or sampling, which generally provide no satisfactory accuracy guarantees. In this work, we propose sidestepping intractable inference altogether by learning ensembles of tractable sub-models as part of a structured prediction cascade. We focus in particular on problems with high-treewidth and large state-spaces, which occur in many computer vision tasks. Unlike other variational methods, our ensembles do not enforce agreement between sub-models, but filter the space of possible outputs by simply adding and thresholding the max-marginals of each constituent model. Our framework jointly estimates parameters for all models in the ensemble for each level of the cascade by minimizing a novel, convex loss function, yet requires only a linear increase in computation over learning or inference in a single tractable sub-model. We provide a generalization bound on the filtering loss of the ensemble as a theoretical justification of our approach, and we evaluate our method on both synthetic data and the task of estimating articulated human pose from challenging videos. We find that our approach significantly outperforms loopy belief propagation on the synthetic data and a state-of-the-art model on the pose estimation/tracking problem.
dc.description.commentsD. Weiss, B. Sapp, and B. Taskar, Sidestepping Intractable Inference with Structured Ensemble Cascades. ;In Proceedings of NIPS. 2010, 2415-2423.
dc.identifier.urihttps://repository.upenn.edu/handle/20.500.14332/6571
dc.legacy.articleid1547
dc.legacy.fulltexturlhttps://repository.upenn.edu/cgi/viewcontent.cgi?article=1547&context=cis_papers&unstamped=1
dc.source.issue511
dc.source.journalDepartmental Papers (CIS)
dc.source.statuspublished
dc.subject.otherComputer Sciences
dc.titleSidestepping Intractable Inference with Structured Ensemble Cascades
dc.typePresentation
digcom.identifiercis_papers/511
digcom.identifier.contextkey3050732
digcom.identifier.submissionpathcis_papers/511
digcom.typeconference
dspace.entity.typePublication
relation.isAuthorOfPublication48084f74-55a3-43da-96d7-8a01c512b3b9
relation.isAuthorOfPublication48084f74-55a3-43da-96d7-8a01c512b3b9
relation.isAuthorOfPublication.latestForDiscovery48084f74-55a3-43da-96d7-8a01c512b3b9
upenn.schoolDepartmentCenterDepartmental Papers (CIS)
Files
Original bundle
Now showing 1 - 1 of 1
Thumbnail Image
Name:
sec_nips10.pdf
Size:
3.77 MB
Format:
Adobe Portable Document Format
Download
Collection
Presentations