English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Zendy Plus

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Zendy Tools

Zendy Open

Machine learning on encrypted data has received a lot of attention thanks torecent breakthroughs in homomorphic encryption and secure multi-partycomputation. It allows outsourcing computation to untrusted servers withoutsacrificing privacy of sensitive data. We propose a practical framework toperform partially encrypted and privacy-preserving predictions which combinesadversarial training and functional encryption. We first present a newfunctional encryption scheme to efficiently compute quadratic functions so thatthe data owner controls what can be computed but is not involved in thecalculation: it provides a decryption key which allows one to learn a specificfunction evaluation of some encrypted data. We then show how to use it inmachine learning to partially encrypt neural networks with quadratic activationfunctions at evaluation time, and we provide a thorough analysis of theinformation leaks based on indistinguishability of data items of the samelabel. Last, since most encryption schemes cannot deal with the lastthresholding operation used for classification, we propose a training method toprevent selected sensitive features from leaking, which adversarially optimizesthe network against an adversary trying to identify these features. This isinteresting for several existing works using partially encrypted machinelearning as it comes with little reduction on the model's accuracy andsignificantly improves data privacy.

Partially Encrypted Machine Learning using Functional Encryption