Xiaomi Patent Application for CSI Reporting with AI/ML

ChangeBridge: Patent Apps - Networking (H04L)

Published March 26th, 2026

Detected March 26th, 2026

Summary

The USPTO has published a patent application from Beijing Xiaomi Mobile Software Co., Ltd. for a method and apparatus related to CSI reporting using AI/ML models. The application, filed on August 12, 2022, describes a terminal device determining quantization parameters for an AI/ML model used to encode CSI.

View original document View source feed page

What changed

This document is a publication of a patent application by Beijing Xiaomi Mobile Software Co., Ltd. (Application No. 19103379, Publication No. US20260088870A1) filed on August 12, 2022, and published on March 26, 2026. The application details a method for reporting Channel State Information (CSI) using artificial intelligence (AI) and machine learning (ML) models, focusing on determining quantization parameters for encoding CSI.

This is a patent application, not a regulatory rule or guidance. It does not impose any immediate compliance obligations or penalties on regulated entities. Companies operating in the telecommunications and networking sectors, particularly those involved in wireless communication technologies and AI/ML integration, may find the disclosed technology relevant for future product development or intellectual property considerations.

Source document (simplified)

← USPTO Patent Applications

CSI REPORTING METHOD AND APPARATUS, DEVICE, AND SYSTEM

Application US20260088870A1 Kind: A1 Mar 26, 2026

Assignee

Beijing Xiaomi Mobile Software Co., Ltd.

Inventors

Min LIU

Abstract

Disclosed are a method and apparatus for reporting CSI, devices, and a system, relating to the technical field of communications. The method includes: determining, by a terminal device, a quantization parameter corresponding to a CSI encoding model, where the CSI encoding model is an artificial intelligence (AI)/machine learning (ML) model used to encode CSI, and the quantization parameter is used to quantify the encoded CSI obtained by the CSI encoding model.