Home
Deutsch
Advanced Search
Browse by ...
Publication overview
Contact persons
Help
DOI
 |
|
Open Access Green as soon as Postprint is submitted to ZB.
Gradient Self-alignment in Private Deep Learning.
In: (26th International Conference on Medical Image Computing and Computer-Assisted Intervention , MICCAI 2023, 8 - 12 October 2023, Vancouver, CANADA). Berlin [u.a.]: Springer, 2023. 89-97 (Lect. Notes Comput. Sc. ; 14393)
Differential Privacy (DP) has become a gold-standard to preserve privacy in deep learning. Intuitively speaking, DP ensures that the output of a model is approximately invariant to the inclusion or exclusion of a single individual’s data from the training set. There is, however, a trade-off between privacy and utility. DP models tend to perform worse than non-DP models trained on the same data. This is caused by the clipping of per-sample gradients and the addition of noise required for DP guarantees causing an obfuscation of the individual data point’s contribution. In this work, we propose a method to reduce this discrepancy by improving the alignment between the per-sample gradients of each individual training sample with its non-DP gradient by increasing their cosine similarity. Optimizing the alignment in only a relevant subset of gradient dimensions, further improves the performance. We evaluate our method on CIFAR-10 and a pediatric pneumonia chest x-ray dataset.
Altmetric
Additional Metrics?
Edit extra informations
Login
Publication type
Article: Conference contribution
Keywords
Differential Privacy ; Gradient Alignment ; Private Learning
ISSN (print) / ISBN
0302-9743
e-ISSN
1611-3349
Conference Title
26th International Conference on Medical Image Computing and Computer-Assisted Intervention , MICCAI 2023
Conference Date
8 - 12 October 2023
Conference Location
Vancouver, CANADA
Quellenangaben
Volume: 14393,
Pages: 89-97
Publisher
Springer
Publishing Place
Berlin [u.a.]
Non-patent literature
Publications
Institute(s)
Institute for Machine Learning in Biomed Imaging (IML)