English
 
Help Privacy Policy Disclaimer
  Advanced SearchBrowse

Item

ITEM ACTIONSEXPORT
  Counterfactuals uncover the modular structure of deep generative models

Besserve, M., Mehrjou, A., Sun, R., & Schölkopf, B. (2020). Counterfactuals uncover the modular structure of deep generative models. In Eighth International Conference on Learning Representations (ICLR 2020).

Item is

Basic

show hide
Genre: Conference Paper

Files

show Files

Locators

show
hide
Description:
-

Creators

show
hide
 Creators:
Besserve, M1, 2, Author              
Mehrjou, A.3, Author              
Sun, R, Author
Schölkopf, B3, Author              
Affiliations:
1Department Physiology of Cognitive Processes, Max Planck Institute for Biological Cybernetics, Max Planck Society, ou_1497798              
2Max Planck Institute for Biological Cybernetics, Max Planck Society, Spemannstrasse 38, 72076 Tübingen, DE, ou_1497794              
3Dept. Empirical Inference, Max Planck Institute for Intelligent Systems, Max Planck Society, ou_1497647              

Content

show
hide
Free keywords: -
 Abstract: Deep generative models can emulate the perceptual properties of complex image datasets, providing a latent representation of the data. However, manipulating such representation to perform meaningful and controllable transformations in the data space remains challenging without some form of supervision. While previous work has focused on exploiting statistical independence to \textit{disentangle} latent factors, we argue that such requirement can be advantageously relaxed and propose instead a non-statistical framework that relies on identifying a modular organization of the network, based on counterfactual manipulations. Our experiments support that modularity between groups of channels is achieved to a certain degree on a variety of generative models. This allowed the design of targeted interventions on complex image datasets, opening the way to applications such as computationally efficient style transfer and the automated assessment of robustness to contextual changes in pattern recognition systems.

Details

show
hide
Language(s):
 Dates: 2020-04
 Publication Status: Published online
 Pages: -
 Publishing info: -
 Table of Contents: -
 Rev. Type: -
 Identifiers: -
 Degree: -

Event

show
hide
Title: Eighth International Conference on Learning Representations (ICLR 2020)
Place of Event: Addis Ababa, Ethiopia
Start-/End Date: 2020-04-26 - 2020-04-30

Legal Case

show

Project information

show

Source 1

show
hide
Title: Eighth International Conference on Learning Representations (ICLR 2020)
Source Genre: Proceedings
 Creator(s):
Affiliations:
Publ. Info: -
Pages: - Volume / Issue: - Sequence Number: - Start / End Page: - Identifier: -