CryptoNumerics breaks through Data Barriers by using a combination of provably secure cryptographic protocols and numerical methods.
Decentralized Differential Privacy
Make data Differentially Private without injecting noise by applying random sampling followed by optimal attribute generalization with configurable metrics of information loss. Protects the PII in either a single dataset or across decentralized data aligned by CN-Fuse when augmenting either features or observations.
Private Set Intersection
Securely identify the intersection or symmetric differences of elements across sensitive datasets without revealing anything else and without using a “trusted” third party. Align datasets for use in secure decentralized machine learning in CN-Insight when augmenting either features or observations.
Decentralized Secure Machine Learning
Build secure machine learning models across decentralized datasets using techniques from Secure Multiparty Computation without exposing or moving the data. Retain control over the types of models trained on your data and how often.