Interpretable and Differentially Private Predictions
By Frederik Harder, Matthias Bauer, Mijung Park. 2019
63 citations recorded by the source corpus.
Interpretable predictions, which clarify why a machine learning model makes a particular decision, can compromise privacy by revealing the characteristics of individual data points. This raises the central question addressed in this paper: Can models be interpretable without compromising privacy? For complex “big” data fit by correspondingly rich models, balancing privacy and explainability is particularly challenging, such that this question has remained largely unexplored. In this paper, we propose a family of simple models with the aim of approximating complex models using several locally linear maps per class to provide high classification accuracy, as well as differentially private explanations on the classification. We illustrate the usefulness of our approach on several image benchmark datasets as well as a medical dataset.
Subfields: Artificial Intelligence
How researchers see this paper
The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.
In Artificial Intelligence
- Adrian Weller
- Umang Bhatt
- José M. F. Moura
- Nguyen Binh Truong
- Tarek R. Besold
- Roberto Confalonieri
- Kai Sun
- Siyao Wang
Citations in the corpus
- Explainable machine learning in deployment (2019, methodology): “[25] develop a methodology for training a differentially private model that generates local and global explanations”
- Secure, privacy-preserving and federated machine learning in medical imaging (2020, background): “Current research about interpretable private algorithms 93 can alleviate this issue.”
- A historical perspective of explainable Artificial Intelligence (2020, methodology): “Harder, Bauer, and Park (2020) is an example of methods of this kind.”
- Privacy preservation in federated learning: An insightful survey from the GDPR perspective (2020, methodology): “omising solution to comply with this GDPR principle is to design a new type of ML models with associated algorithms that are inherently interpretable, which encourages responsible ML governance [124]–[127]. 10https://ico.org.uk/for-organisations/guide-to-data-protection/guide-tothe-general-data-prot”