Explaining classifiers by constructing familiar concepts

Details

Ressource 1Request a copy Under indefinite embargo.
UNIL restricted access
State: Public
Version: author
License: CC BY 4.0
Serval ID
serval:BIB_B0149200E7B1
Type
Article: article from journal or magazin.
Collection
Publications
Institution
Title
Explaining classifiers by constructing familiar concepts
Journal
Machine Learning
Author(s)
Schneider Johannes, Vlachos Michalis
ISSN
0885-6125
1573-0565
Publication state
Published
Issued date
25/03/2022
Peer-reviewed
Oui
Language
english
Abstract
Interpreting a large number of neurons in deep learning is difficult. Our proposed ‘CLAssi- fier-DECoder’ architecture (ClaDec) facilitates the understanding of the output of an arbi- trary layer of neurons or subsets thereof. It uses a decoder that transforms the incompre- hensible representation of the given neurons to a representation that is more similar to the domain a human is familiar with.
Keywords
Artificial Intelligence, Software
Open Access
Yes
Create date
29/04/2022 17:13
Last modification date
07/09/2022 6:39
Usage data