Sensor compensation using backpropagation - IBM patent US12596930B2

ChangeBridge: Patent Grants - AI & Computing (G06N)

Published April 7th, 2026

Detected April 7th, 2026

Summary

USPTO granted IBM patent US12596930B2 on April 7, 2026. The patent covers training convolutional neural networks (CNNs) and modifying sensor data using compensation values derived from error calculations. Assignee is International Business Machines Corporation with inventors Minsik Cho, Inseok Hwang, and Chungkuk Yoo. The 18-claim patent relates to sensor compensation in machine learning systems.

View original document View source feed page

What changed

USPTO granted IBM patent US12596930B2 for sensor compensation using backpropagation in convolutional neural networks. The patent describes training first and second CNNs using training images, adding an interface layer to process sensor data inputs, and iteratively modifying compensation values based on error calculations between result vectors.

Technology companies developing neural network-based sensor systems should review this patent landscape for potential licensing considerations. Patent holders may enforce exclusive rights against entities practicing similar sensor compensation methods using CNNs and backpropagation techniques.

What to do next

Monitor for updates

Source document (simplified)

← USPTO Patent Grants

Sensor compensation using backpropagation

Grant US12596930B2 Kind: B2 Apr 07, 2026

Assignee

INTERNATIONAL BUSINESS MACHINES CORPORATION

Inventors

Minsik Cho, Inseok Hwang, Chungkuk Yoo

Abstract

An embodiment includes training a first convolutional neural network (CNN) using a plurality of training images to generate first and second trained CNNs, and then adding an interface layer to the second trained CNN. The embodiment processes a first and second images in a sequence of images using the first trained CNN to generate a first and second result vectors. The embodiment also processes the second image using the second trained CNN and sensor data input to the interface layer to generate a third result vector. The embodiment modifies the sensor data using a compensation value. The embodiment compares the third result vector to the second result vector to generate an error value, and then calculates a modified compensation value using the error value. The embodiment then generates a sensor-compensated trained CNN based on the second trained CNN with the modified compensation value.