Data Reconstruction Using Machine-Learning Predictive Coding

ChangeBridge: Patent Apps - AI & Computing (G06N)

Published March 26th, 2026

Detected March 31st, 2026

Summary

USPTO published patent application US20260087314A1 for a machine-learning method that reconstructs data samples in a time series using predictive coding. The method generates reconstructed versions of first and second data samples, then uses a neural network to predict intermediate data samples positioned between them. The application (No. 19107781) was filed July 27, 2023 and published March 26, 2026.

View original document View source feed page

What changed

USPTO published patent application US20260087314A1 titled 'Data Reconstruction Using Machine-Learning Predictive Coding.' The invention describes a method for generating reconstructed data samples corresponding to versions of first and second data samples in a time series, then providing these reconstructed samples as inputs to a neural network configured for machine-learning predictive coding. The network generates network-predicted data samples corresponding to predicted versions of particular data samples positioned between the first and second data samples. CPC classifications are G06N 3/0455 and G06N 3/044. Inventors are Guillaume Konrad Sautiere, Vivek Rajendran, and Zisis Iason Skordilis.

Patent applications do not create compliance obligations for third parties. Technology companies and AI developers researching predictive modeling or neural network applications should review the published claims to assess potential impacts on freedom-to-operate or competitive landscape. Investors in AI and machine-learning technology may find this application relevant for competitive analysis or patent portfolio evaluation.

Source document (simplified)

← USPTO Patent Applications

DATA RECONSTRUCTION USING MACHINE-LEARNING PREDICTIVE CODING

Application US20260087314A1 Kind: A1 Mar 26, 2026

Inventors

Guillaume Konrad SAUTIERE, Vivek RAJENDRAN, Zisis Iason SKORDILIS

Abstract

A method includes generating a first reconstructed data sample corresponding to a reconstructed version of a first data sample in a time series of data samples. The method includes generating a second reconstructed data sample corresponding to a reconstructed version of a second data sample in the time series of data samples. The method includes providing the first reconstructed data and the second reconstructed data sample as inputs to a neural network. The neural network is configured to use machine-learning predictive coding to generate a network-predicted data sample. The network-predicted data sample corresponds to a predicted version of a particular data sample in the time series of data samples that is positioned between the first data sample and the second data sample.