Information processing system with multiple partial reservoirs trained using differential teaching data

ChangeBridge: Patent Grants - AI & Computing (G06N)

Published April 7th, 2026

Detected April 7th, 2026

Summary

USPTO granted Fujitsu Limited Patent US12596927B2 covering an information processing system with multiple partial reservoirs trained using differential teaching data. The patent describes a reservoir computing architecture with common input layer, first and second output layers, and partial reservoirs of varying sizes for optimized training. The patent includes 7 claims and is classified under CPC G06N 3/045, G06N 3/065, and G06N 3/08 (machine learning/neural networks).

View original document View source feed page

What changed

USPTO issued Patent US12596927B2 to Fujitsu Limited for an information processing system using multiple partial reservoirs trained with differential teaching data. The system includes a common input layer, first and second output layers outputting readout values based on inputs, a first partial reservoir with input and first output layers, and a second larger partial reservoir with size between the input and second output layers. Training calculates third and fourth output weights to reduce differences between product sum values and teaching data.

This patent grant establishes enforceable intellectual property rights in reservoir computing and neural network training methods under CPC classifications G06N 3/045, G06N 3/065, and G06N 3/08. Technology companies developing machine learning systems, particularly those using reservoir computing architectures, should monitor for potential licensing requirements or design-around considerations to avoid infringement on Fujitsu's protected methods.

What to do next

Monitor for updates

Source document (simplified)

← USPTO Patent Grants

Information processing system having an information processing device and multiple partial reservoirs that are trained, information processing device that trains multiple partial reservoirs, and non-transitory computer readable memory medium that stores information processing program for training multiple partial reservoirs

Grant US12596927B2 Kind: B2 Apr 07, 2026

Assignee

Fujitsu Limited

Inventors

Shoichi Miyahara

Abstract

A reservoir includes a common input layer, first and second output layers that outputs a first and a second readout values based on an input, a first partial reservoir including the input layer and the first output layer, and a second partial reservoir having a size between the input layer and the second output layer larger than the size of the first partial reservoir, and the training processing including: first calculating a third output weight that reduces a difference between a first product sum value of a third readout value and a first output weight; and second calculating a fourth output weight that reduces a difference between a second product sum value of a fourth readout value and a second output weight and differential teaching data that is a difference between a third product sum value of the third readout value and the third output weight and the teaching data.