Controlling Directions Orthogonal to a Classifier
Name
Xu-ylxu-SM-EECS-2022-thesis.pdf
Description
Thesis PDF
Size
2.14 MB
Format
Adobe PDF
Checksum (MD5)
ea768f711781512f1631b015d52edcee
Author(s)
Xu, Yilun
Advisor(s)
Jaakkola, Tommi
Date Issued
September 2022
Publisher
Massachusetts Institute of Technology
Abstract
We propose to identify directions invariant to a given classifier so that these directions can be controlled in tasks such as style transfer. While orthogonal decomposition is directly identifiable when the given classifier is linear, we formally define a notion of orthogonality in the non-linear case. We also provide a surprisingly simple method for constructing the orthogonal classifier (a classifier utilizing directions other than those of the given classifier). Empirically, we present three use cases where controlling orthogonal variation is important: style transfer, domain adaptation, and fairness. The orthogonal classifier enables desired style transfer when domains vary in multiple aspects, improves domain adaptation with label shifts and mitigates the unfairness as a predictor.
MIT Department
Massachusetts Institute of Technology. Department of Electrical Engineering and Computer Science
Terms of Use
In Copyright - Educational Use Permitted
Copyright MIT
Persistent DSpace Link
