When Differential Privacy Meets Interpretability: A Case Study
arXiv:2106.13203
Abstract
Given the increase in the use of personal data for training Deep Neural Networks (DNNs) in tasks such as medical imaging and diagnosis, differentially private training of DNNs is surging in importance and there is a large body of work focusing on providing better privacy-utility trade-off. However, little attention is given to the interpretability of these models, and how the application of DP affects the quality of interpretations. We propose an extensive study into the effects of DP training on DNNs, especially on medical imaging applications, on the APTOS dataset.
arXiv comment4 pages, 7 figures; Extended abstract presented at RCV-CVPR'21
References in corpus (7)
- Deep Learning with Differential Privacy
- Advancements in Image Classification using Convolutional Neural Network
- Learning Deep Features for Discriminative Localization
- Privacy in Deep Learning: A Survey
- Improving the Gaussian Mechanism for Differential Privacy: Analytical Calibration and Optimal Denoising
- Local Differential Privacy: a tutorial
- FedPandemic: A Cross-Device Federated Learning Approach Towards Elementary Prognosis of Diseases During a Pandemic