WEIGHT PROCESSING FOR A NEURAL NETWORK

Organization Name

Inventor(s)

John Wakefield Brothers, Iii of Calistoga CA (US)

WEIGHT PROCESSING FOR A NEURAL NETWORK - A simplified explanation of the abstract

This abstract first appeared for US patent application 20240048152 titled 'WEIGHT PROCESSING FOR A NEURAL NETWORK

Simplified Explanation

The abstract describes a system and method for processing data for a neural network. The system includes memory to receive data bits defining a kernel of weights and a data processing unit. The data processing unit receives bits defining a kernel of weights for the neural network, generates a set of mask bits indicating the presence of zero-valued or non-zero value weights in the kernel, and transmits the non-zero value weights and the set of mask bits for storage.

The system processes data for a neural network.
It uses memory to store data bits defining a kernel of weights.
The data processing unit receives bits defining a kernel of weights.
It generates a set of mask bits indicating the presence of zero-valued or non-zero value weights in the kernel.
The non-zero value weights and the set of mask bits are transmitted for storage.

Potential applications of this technology:

Neural network training and inference
Image recognition and classification
Natural language processing
Speech recognition
Autonomous vehicles
Robotics

Problems solved by this technology:

Efficient storage and processing of neural network weights
Reduction of memory requirements
Optimization of neural network performance

Benefits of this technology:

Improved efficiency in neural network processing
Faster training and inference times
Reduced memory footprint
Enhanced performance and accuracy in various applications

Original Abstract Submitted

systems and methods for processing data for a neural network are described. the system comprises non-transitory memory configured to receive data bits defining a kernel of weights, the data bits being suitable for processing input data; and a data processing unit, configured to: receive bits defining a kernel of weights for the neural network, the kernel of weights comprising one or more non-zero value weights and one or more zero-valued weights; generate a set of mask bits, a position of each bit in the set of mask bits corresponds to a position within the kernel of weights and the value of each bit indicates whether a weight in the corresponding position is a zero-valued weight or a non-zero value weight; and transmit the non-zero value weights and the set of mask bits for storage, the non-zero value weights and the set of mask bits represent the kernel of weights.