Title: | Hyperparameters and neural architectures in differentially private deep learning |
Author: | Tobaben, Marlon |
Other contributor: |
Helsingin yliopisto, Matemaattis-luonnontieteellinen tiedekunta
University of Helsinki, Faculty of Science Helsingfors universitet, Matematisk-naturvetenskapliga fakulteten |
Publisher: | Helsingin yliopisto |
Date: | 2022 |
Language: | eng |
URI: |
http://urn.fi/URN:NBN:fi:hulib-202206222946
http://hdl.handle.net/10138/345424 |
Thesis level: | master's thesis |
Degree program: |
Datatieteen maisteriohjelma
Master's Programme in Data Science Magisterprogrammet i data science |
Specialisation: |
ei opintosuuntaa
no specialization ingen studieinriktning |
Abstract: | Using machine learning to improve health care has gained popularity. However, most research in machine learning for health has ignored privacy attacks against the models. Differential privacy (DP) is the state-of-the-art concept for protecting individuals' data from privacy attacks. Using optimization algorithms such as the DP stochastic gradient descent (DP-SGD), one can train deep learning models under DP guarantees. This thesis analyzes the impact of changes to the hyperparameters and the neural architecture on the utility/privacy tradeoff, the main tradeoff in DP, for models trained on the MIMIC-III dataset. The analyzed hyperparameters are the noise multiplier, clipping bound, and batch size. The experiments examine neural architecture changes regarding the depth and width of the model, activation functions, and group normalization. The thesis reports the impact of the individual changes independently of other factors using Bayesian optimization and thus overcomes the limitations of earlier work. For the analyzed models, the utility is more sensitive to changes to the clipping bound than to the other two hyperparameters. Furthermore, the privacy/utility tradeoff does not improve when allowing for more training runtime. The changes to the width and depth of the model have a higher impact than other modifications of the neural architecture. Finally, the thesis discusses the impact of the findings and limitations of the experiment design and recommends directions for future work. |
Subject: |
differential privacy
deep learning hyperparameter optimization Bayesian optimization |
Total number of downloads: Loading...
Files | Size | Format | View |
---|---|---|---|
Tobaben_Marlon_thesis_2022.pdf | 2.817Mb |
View/ |