On the Robustness of Convolutional Neural Networks to Internal Architecture and Weight Perturbations
Deep convolutional neural networks are generally regarded as robust function approximators. So far, this intuition is based on perturbations to external stimuli such as the images to be classified. Here we explore the robustness of convolutional neural networks to perturbations to the internal weigh...
Main Authors: | Cheney, Nicholas, Schrimpf, Martin, Kreiman, Gabriel |
---|---|
Format: | Technical Report |
Language: | en_US |
Published: |
Center for Brains, Minds and Machines (CBMM), arXiv
2017
|
Online Access: | http://hdl.handle.net/1721.1/107935 |
Similar Items
-
Towards Certificated Model Robustness Against Weight Perturbations
by: Weng, Tsui-Wei, et al.
Published: (2022) -
Convolutional neural network architectures for predicting DNA–protein binding
by: Zeng, Haoyang, et al.
Published: (2017) -
On the robustness of graph neural diffusion to topology perturbations
by: Song, Yang, et al.
Published: (2023) -
Improving the robustness of machine learning system through convolutional neural network
by: Chia, Daryl Jing
Published: (2019) -
CNN-Cert: An Efficient Framework for Certifying Robustness of Convolutional Neural Networks
by: Boopathy, Akhilan, et al.
Published: (2021)