US Patent Application 18221089. ELECTRONIC DEVICE AND METHOD FOR ACCELERATING NEURAL NETWORK COMPUTATIONS simplified abstract
Contents
ELECTRONIC DEVICE AND METHOD FOR ACCELERATING NEURAL NETWORK COMPUTATIONS
Organization Name
Inventor(s)
Hongxiang Fan of Chertsey (GB)
Chun Pong Chau of Chertsey (GB)
Stylianos Venieris of Chertsey (GB)
Alexandros Kouris of Chertsey (GB)
Mohamed S. Abdelfattah of Chertsey (GB)
ELECTRONIC DEVICE AND METHOD FOR ACCELERATING NEURAL NETWORK COMPUTATIONS - A simplified explanation of the abstract
This abstract first appeared for US patent application 18221089 titled 'ELECTRONIC DEVICE AND METHOD FOR ACCELERATING NEURAL NETWORK COMPUTATIONS
Simplified Explanation
The abstract describes an electronic device designed to accelerate machine learning model computations.
- The device includes a first processor that generates query, key, and value matrices using Fast Fourier Transform and butterfly linear transform on input matrices.
- The device also includes a second processor that performs matrix multiplications and softmax operations on the generated matrices.
- The second processor then performs another matrix multiplication using the result of the softmax operation and the value matrix.
- The purpose of this device is to improve the speed and efficiency of machine learning computations.
Original Abstract Submitted
Broadly speaking, the present disclosure generally relate to an electronic device for accelerating machine learning, ML, model computations is provided. The electronic device comprises: a first processor configured to: generate a query matrix, a key matrix, and a value matrix by performing Fast Fourier Transform (FFT) and butterfly linear transform on at least one input matrix, and a second processor configured to: perform a first matrix multiplication between the query matrix and the key matrix, perform a softmax operation on the result of the first matrix multiplication, and perform a second matrix multiplication between the result of the softmax operation and the value matrix.