Publication:
Learning with Aggregate Data

Simple item page
dc.contributor.advisorDaniel Sheldon
dc.contributor.authorSUN, TAO
dc.contributor.departmentUniversity of Massachusetts Amherst
dc.date2024-03-27T17:40:16.000
dc.date.accessioned2024-04-26T15:28:50Z
dc.date.available2024-04-26T15:28:50Z
dc.date.submittedFebruary
dc.date.submitted2019
dc.description.abstractVarious real-world applications involve directly dealing with aggregate data. In this work, we study Learning with Aggregate Data from several perspectives and try to address their combinatorial challenges. At first, we study the problem of learning in Collective Graphical Models (CGMs), where only noisy aggregate observations are available. Inference in CGMs is NP- hard and we proposed an approximate inference algorithm. By solving the inference problems, we are empowered to build large-scale bird migration models, and models for human mobility under the differential privacy setting. Secondly, we consider problems given bags of instances and bag-level aggregate supervisions. Specifically, we study the US presidential election and try to build a model to understand the voting preferences of either individuals or demographic groups. The data consists of characteristic individuals from the US Census as well as voting tallies for each voting precinct. We proposed a fully probabilistic Learning with Label Proportions (LLPs) model with exact inference to build an instance-level model. Thirdly, we study distribution regression. It has similar problem setting to LLPs but builds bag-level models. We experimentally evaluated different algorithms on three tasks, and identified key factors in problem settings that impact the choice of algorithm.
dc.description.degreeDoctor of Philosophy (PhD)
dc.description.departmentComputer Science
dc.identifier.doihttps://doi.org/10.7275/13489220
dc.identifier.orcidN/A
dc.identifier.urihttps://hdl.handle.net/20.500.14394/17706
dc.relation.urlhttps://scholarworks.umass.edu/cgi/viewcontent.cgi?article=2572&context=dissertations_2&unstamped=1
dc.source.statuspublished
dc.subjectmachine learning
dc.subjectgraphical models
dc.subjectArtificial Intelligence and Robotics
dc.titleLearning with Aggregate Data
dc.typeopenaccess
dc.typearticle
dc.typedissertation
digcom.contributor.authorisAuthorOfPublication|email:suntao.st@gmail.com|institution:University of Massachusetts Amherst|SUN, TAO
digcom.identifierdissertations_2/1483
digcom.identifier.contextkey13489220
digcom.identifier.submissionpathdissertations_2/1483
dspace.entity.typePublication
Files
Original bundle
Now showing 1 - 1 of 1
No Thumbnail Available
Name:
tao.pdf
Size:
4.78 MB
Format:
Adobe Portable Document Format
Name: tao.pdf
Size: 4.78 MB
Kind: Adobe PDF
Collections
Dissertations