US Patent Issued on Sept. 8 for "Label inference in split learning defenses" (California Inventors)

ALEXANDRIA, Va., Sept. 8 -- United States Patent no. 12,731,035, issued on Sept. 8.

"Label inference in split learning defenses" was invented by Shangyu Xie (Los Angeles), Jiankai Sun (Los Angeles), Xin Yang (Los Angeles), Yuanshun Yao (Los Angeles), Tianyi Liu (Los Angeles) and Taiqing Wang (Los Angeles).

According to the abstract* released by the U.S. Patent & Trademark Office: "Split learning is provided to train a composite neural network (CNN) model that is split into first and second submodels, including receiving a noise-laden backpropagation gradient, training the surrogate submodel by optimizing a gradient distance loss, and computing an updated dummy label using the first submodel and the trained surrogate submodel to infer labe...