Parameter optimization method, electronic device, and storage medium

A non-transitory computer-readable storage medium storing computer instructions for execution by at least one processor to:

• obtain a Kronecker factor matrix of a higher-order information matrix of a neural network model, wherein the higher-order information matrix corrects a first-order gradient of the neural network model, and wherein the Kronecker factor matrix is one of a pair of Kronecker factor matrices that indicate the higher-order information matrix;
• segment the Kronecker factor matrix to obtain a plurality of square matrices, wherein the plurality of square matrices are submatrices of the Kronecker factor matrix, and main diagonal lines of the plurality of square matrices each are in a one-to-one correspondence with a part of a main diagonal line of the Kronecker factor matrix;
• and
• adjust a parameter of the neural network model based on the plurality of square matrices.

A parameter optimization apparatus, comprising:

• at least one processor;
• and
• one or more memories coupled to the at least one processor and storing programming instructions for execution by the at least one processor to:
• obtain a Kronecker factor matrix of a higher-order information matrix of a neural network model, wherein the higher-order information matrix corrects a first-order gradient of the neural network model, and wherein the Kronecker factor matrix is one of a pair of Kronecker factor matrices that indicate the higher-order information matrix;
• segment the Kronecker factor matrix to obtain a plurality of square matrices, wherein the plurality of square matrices are submatrices of the Kronecker factor matrix, and main diagonal lines of the plurality of square matrices each are in a one-to-one correspondence with a part of a main diagonal line of the Kronecker factor matrix;
• and
• adjust a parameter of the neural network model based on the plurality of square matrices.