Dear guest, welcome to this publication database. As an anonymous user, you will probably not have edit rights. Also, the collapse status of the topic tree will not be persistent. If you like to have these and other options enabled, you might ask Admin (Ivan Eggel) for a login account.
 [BibTeX] [RIS]
Interpreting intentionally flawed models with linear probes
Type of publication: Inproceedings
Citation:
Booktitle: ICCV workshop on statistical deep learning in computer vision
Year: 2019
Location: Seoul, Korea
Abstract: The representational differences between generalizing networks and intentionally flawed models can be insightful on the dynamics of network training. Do memorizing networks, e.g. networks that learn random label correspondences, focus on specific patterns in the data to memorize the labels? Are the features learned by a generalizing network affected by randomization of the model parameters? In high-risk applications such as medical, legal or financial domains, highlighting the representational differences that help generalization may be even more important than model performance itself. In this paper, we probe the activations of intermediate layers with linear classification and regression. Results show that the bias towards simple solutions of generalizing networks is maintained even when statistical irregularities are intentionally introduced.
Keywords: Deep Learning, interpretability, wrong labels
Authors Graziani, Mara
Müller, Henning
Andrearczyk, Vincent
Added by: []
Total mark: 0
Attachments
  • grazianiSDLCV19.pdf
Notes
    Topics