Device and method for executing artificial neural network self-learning operation

Abstract

The invention discloses a device and method for executing artificial neural network self-learning operation. The device comprises an instruction storage unit, a controller unit, a data access unit, an interconnection module, a primary operation module and a plurality of secondary operation modules. According to the device and method, the self-learning pre-training of a multilayer neural network can adopt a layer-by-layer training manner; for each layer of network, the self-learning pre-training is completed after multiple operations are iterated until the weight is updated to be smaller than a certain threshold value. Each iteration process can be divided into four stages, calculation is carried out in the first three stages to respectively generate a first-order hidden layer median, a first-order visible layer median and a second-order hidden layer median, and the weight is updated in the last stage by utilizing the medians in the first three stages.

Description

technical field [0001] The invention relates to artificial neural network technology, in particular to a device and method for performing artificial neural network self-learning operations. Background technique [0002] Multi-layer artificial neural networks are widely used in the fields of pattern recognition, image processing, function approximation, and optimization calculations. In recent years, multi-layer artificial networks have been favored by academic circles and researchers due to their high recognition accuracy and good parallelism. industry is getting more and more attention. [0003] A typical multi-layer artificial neural network training method is the backpropagation (BP) algorithm. This method is a representative type of supervised learning, which requires a large number of labeled training samples during the training process, but the cost of sample collection is very high. At the same time, during the training process of this method, the error correction s...

AI Technical Summary

Problems solved by technology

[0006] What the present invention is to solve is that in the prior art, general-purpose processors (GPU, CPU) need a series of simple operations and memory access operations for pre-training of multi-layer neural networks, and the power consumption overhead of front-end decoding is large and existing general-purpose processing However, there are many problems such as high data memory access overhead and low computing performance of a single general-purpose processor.

Images