Assessing the usefulness of NetworkDissection in identifying the interpretability of facial characterization networks
Name
1098171971-MIT.pdf
Size
4.44 MB
Format
Adobe PDF
Checksum (MD5)
f93f7065bcad84937d0abd957ab988d5
Author(s)
Dethy, Elizabeth(Elizabeth A.)
Advisor(s)
Daniel J. Weitzner.
Date Issued
2018
Publisher
Massachusetts Institute of Technology
Abstract
In this thesis I test for the emergence of policy relevant features as disentangled, human interpretable representations in facial characterization networks. I probe an age and gender classifier using the NetworkDissection method with a hand labelled dataset of facial features, skin tones, and textures. Facial features and skin tones emerge as disentangled concepts in each of the networks probed. The emergence of these features in a smaller image classification network indicates the effectiveness of the NetworkDissection method in contexts other than ones studied in the original paper. Moreover, the emergence of policy relevant features, skin tones, indicates the method may be effective in identifying policy sensitive attributes. I also analyze the robustness of the NetworkDissection technique itself to changes in a key component of the experimental setup: the source of ground truth human understandable concepts. The results demonstrate the technique reliably applies labels when new concepts and samples are added to the set of ground truth labels.
Description
This electronic version was submitted by the student author. The certified thesis is available in the Institute Archives and Special Collections.
Thesis: M. Eng., Massachusetts Institute of Technology, Department of Electrical Engineering and Computer Science, 2018
Cataloged from student-submitted PDF version of thesis.
Includes bibliographical references (pages 55-56).
Subjects
Electrical Engineering and Computer Science.
MIT Department
Massachusetts Institute of Technology. Department of Electrical Engineering and Computer Science
Terms of Use
MIT theses are protected by copyright. They may be viewed, downloaded, or printed from this source but further reproduction or distribution in any format is prohibited without written permission.
Persistent DSpace Link
