2026-09-18T18:56:45Zhttps://dspace.mit.edu/server/oai/requestoai:dspace.mit.edu:1721.1/1392532022-01-15T03:25:12Zcom_1721.1_7582com_1721.1_7581col_1721.1_131023
Han, Song
Gschwind, Katharina
Massachusetts Institute of Technology. Department of Electrical Engineering and Computer Science
2022-01-14T14:59:39Z
2022-01-14T14:59:39Z
2021-06
2021-06-17T20:13:15.997Z
https://hdl.handle.net/1721.1/139253
Novel machine learning architectures can adeptly learn to predict user response for recommender systems. However, these model architectures are often effective at the cost of large computational, and memory, cost. This limits their ability to run on edge devices with smaller hardwares, such as smartphones, which is a popular use case for recommender systems. We address this issue in this thesis by studying how compression of recommender system models can significantly reduce model computation cost, and edge device runtime, while preserving prediction accuracy. Furthermore, we present a new compression-based AutoML method for feature set generation in architectures which incorporate explicit feature interactions. This works as a tool to build efficient recommender system models, and is applicable to many state of the art model designs. Applying this AutoML shows initial gains in model performance.
M.Eng.
Massachusetts Institute of Technology
In Copyright - Educational Use Permitted
Copyright MIT
http://rightsstatements.org/page/InC-EDU/1.0/
Model Compression and AutoML for Efficient Click-Through Rate Prediction
Thesis
application/pdf
Master
Master of Engineering in Electrical Engineering and Computer Science
Publication
unknown
unknown
<Publication xmlns="https://www.openaire.eu/cerif-profile/1.1/" id="a3871e28-41f8-4145-8536-3c75635d5f00">
<Type xmlns="https://www.openaire.eu/cerif-profile/vocab/COAR_Publication_Types">http://purl.org/coar/resource_type/c_1843</Type>
<Title>Model Compression and AutoML for Efficient Click-Through Rate Prediction</Title>
<PublishedIn>
<Publication>
</Publication>
</PublishedIn>
<PublicationDate>2021-06</PublicationDate>
<Authors>
<Author>
<DisplayName>Gschwind, Katharina</DisplayName>
<Affiliation>
<OrgUnit>
</OrgUnit>
</Affiliation>
</Author>
</Authors>
<Editors>
</Editors>
<Publishers>
<Publisher>
<DisplayName>Massachusetts Institute of Technology</DisplayName>
<OrgUnit />
</Publisher>
</Publishers>
<License>http://rightsstatements.org/page/InC-EDU/1.0/</License>
<Abstract>Novel machine learning architectures can adeptly learn to predict user response for recommender systems. However, these model architectures are often effective at the cost of large computational, and memory, cost. This limits their ability to run on edge devices with smaller hardwares, such as smartphones, which is a popular use case for recommender systems. We address this issue in this thesis by studying how compression of recommender system models can significantly reduce model computation cost, and edge device runtime, while preserving prediction accuracy. Furthermore, we present a new compression-based AutoML method for feature set generation in architectures which incorporate explicit feature interactions. This works as a tool to build efficient recommender system models, and is applicable to many state of the art model designs. Applying this AutoML shows initial gains in model performance.</Abstract>
<Access xmlns="http://purl.org/coar/access_right"
>
</Access>
</Publication>