Show simple item record

dc.contributor.advisorQian, Xiaoning
dc.creatorXiang, Ziyu
dc.date.accessioned2021-05-04T23:06:21Z
dc.date.created2020-12
dc.date.issued2020-11-17
dc.date.submittedDecember 2020
dc.identifier.urihttps://hdl.handle.net/1969.1/192867
dc.description.abstractAutomatic Feature Engineering (AFE) aims to extract useful knowledge for interpretable predictions given data for the machine learning tasks of interest. Here, we develop AFE to extract dependency relationships that can be interpreted with functional formulas in order to discover physics meaning or new hypotheses for the problems of interest. We focus on materials science applications, where interpretable predictive modeling may provide a principled understanding of materials systems and guide new materials discovery. It is often computationally prohibitive to exhaust all the potential relationships to construct and search the whole feature space to identify interpretable and predictive features. We develop and evaluate new AFE strategies by exploring a feature generation tree with deep Q-network for scalable and efficient exploration policies. The developed reinforcement learning based AFE strategies are benchmarked with the existing AFE methods on several materials science datasets.en
dc.format.mimetypeapplication/pdf
dc.language.isoen
dc.subjectFeature Engineeringen
dc.subjectAutomated Feature Engineeringen
dc.subjectFeature Representationen
dc.subjectReinforcement Learningen
dc.subjectMachine Learningen
dc.titleAutomated Feature Engineering with Reinforcement Learning
dc.typeThesisen
thesis.degree.departmentElectrical and Computer Engineeringen
thesis.degree.disciplineElectrical Engineeringen
thesis.degree.grantorTexas A&M Universityen
thesis.degree.nameMaster of Scienceen
thesis.degree.levelMastersen
dc.contributor.committeeMemberKalathil, Dileep
dc.contributor.committeeMemberLiu, Tie
dc.contributor.committeeMemberHu, Xia
dc.type.materialtexten
dc.date.updated2021-05-04T23:06:22Z
local.embargo.terms2022-12-01
local.embargo.lift2022-12-01
local.etdauthor.orcid0000-0003-0925-8705


Files in this item

Thumbnail

This item appears in the following Collection(s)

Show simple item record