Neural network optimization circuit

CN119026653AActive Publication Date: 2024-11-26INST OF AUTOMATION CHINESE ACAD OF SCI

Patent Information

Application Number
CN202411117434.X
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-08-14
Publication Date
2024-11-26
Estimated Expiration
2044-08-14

Smart Images

  • Figure CN119026653A_ABST
    Figure CN119026653A_ABST
Patent Text Reader

Abstract

The invention relates to the technical field of circuits, and discloses a neural network optimization circuit which trains multiple pieces of target training data at the same time, carries out error optimization on the multiple pieces of target training data at the same time, is high in parallelism and high in calculation speed, has on-chip learning ability, does not have discontinuous signals such as switching signals and clock signals in the training process, and is high in reliability. And rapid convergence is realized under the guarantee of a strict mathematical theory, analog circuits are completely used, and the expansion cost is low. Moreover, the second differentiator is post-positioned to the output position of the adder from the disturbance source, so that the original convergence direction of the circuit can be kept, the requirement on disturbance power is reduced, and the stability of the network is improved.
Need to check novelty before this filing date? Find Prior Art

Citation Information

Patent Citations

  • A design method of a linear equation solver based on a variable parameter convergent neural network

    CN109033021A

  • Group Lasso-based neural network cutting method for power amplifier

    CN110414565A

  • Hardware accelerator and hardware accelerator method

    CN115374916A

  • Interpretable Neural Networks for Nonlinear Control

    US20230117488A1

Cited By

  • Neural network optimization circuit

    WO2026037447A1