IBM Senior Research Scientist: IBM has completed field trials in the field of fully homomorphic encryption

According to arstechnica.com, IBM senior research scientist Flavio Bergamaschi stated that IBM has successfully completed field trials in the field of Fully Homomorphic Encryption (FHE). IBM signed a contract with a large U.S. bank and a large European bank to use real-time data to conduct field trials of FHE technology. Among them, the FHE experiment on Bank of America is based on a machine learning model to predict the possibility of loan issuance, and provides a large number of customers with retail banking, loan and investment information. The data set consists of 364,000 items and has hundreds of feature points. The model needs to identify relatively rare events (about 1%) in the data set. The success of the experiment was demonstrated by obtaining encrypted data predictions with similar accuracy equivalent to the base, variable selection with similar accuracy, and calculation costs within an acceptable range. Note: FHE is an encryption algorithm that can protect data privacy. It allows direct mathematical operations on encrypted data, and the result is still correct after decryption. In 2009, IBM’s Craig Gentry first proposed a fully homomorphic based on ideal lattice algorithm.