TRUST-REGION AWARE NEURAL NETWORK ARCHITECTURE SEARCH FOR KNOWLEDGE DISTILLATION

    公开(公告)号:US20230153577A1

    公开(公告)日:2023-05-18

    申请号:US17986803

    申请日:2022-11-14

    CPC classification number: G06N3/0454

    Abstract: A processor-implemented method of searching for a neural network architecture includes defining a search space of student neural network architectures for knowledge distillation. The search space includes multiple convolutional operators and multiple transformer operators. A trust-region Bayesian optimization is performed to select a student neural network architecture from the search space based on a pre-defined teacher model.

Patent Agency Ranking