Methods and arrangements to quantize a neural network with machine learning

US11216719B2 · kind B2 · utility

Assignee

• Intel Corporation · US

Inventors

• Somdeb Majumdar · San Diego, US
• Ron Banner · Haifa, IL
• Marcel Nassar · San Diego, US
• Lior Storfer · Nes Ziona, IL
• Adnan Agbaria · Haifa, IL
• Evren Tumer · Temecula, US
• Tristan Webb · San Diego, US
• Xin Wang · Woodridge, US

Key dates

• Filing date: Jun 15, 2018
• Grant date: Jan 4, 2022
• Expiry date: Aug 3, 2040

Classification

• Technology area (CPC G): Physics
• CPC primary: G06N3/065
• WIPO field: Computer technology
• WIPO sector: Electrical engineering

Abstract

Logic may quantize a primary neural network. Logic may generate, by a secondary neural network logic circuitry for a primary neural network logic circuitry, quantization parameters. The primary neural network logic circuitry may comprise a primary neural network with multiple layers trainable with an objective function. Each of the multiple layers of the primary neural network may comprise multiple tensors. The secondary neural network logic circuitry may comprise one or more secondary neural networks trainable with the objective function to output the quantization parameters to the tensors.