Semantic source coding method and apparatus

The method addresses the challenges of mobile communication systems by using multimodal and unimodal learning models to extract and encode data from various modalities, enabling efficient and accurate semantic communication that ensures downstream tasks are performed as intended.

WO2024228411A1PCT designated stage expired Publication Date: 2024-11-07LG ELECTRONICS INC
View PDF 1 Cites 0 Cited by

Patent Information

Application Number
PCT/KR2023/006128
Authority / Receiving Office
WO · WO
Patent Type
Applications
Current Assignee / Owner
Filing Date
2023-05-04
Publication Date
2024-11-07

AI Technical Summary

Technical Problem

Current mobile communication systems face challenges in supporting explosive data traffic, high transmission rates, a large number of connected devices, very low end-to-end latency, and high energy efficiency, while also struggling with accurately conveying the intended meaning in semantic communication due to limitations in symbol transmission and semantics.

Method used

The method employs multimodal and unimodal learning models to enable semantic communication by extracting raw data from various modalities, performing encoding based on data modality-specific encoders, and generating semantic representation packets that can be transmitted and updated to ensure downstream tasks operate according to the source's intent.

Benefits of technology

This approach allows for effective semantic communication that supports downstream tasks by generating rich semantic expressions and updating background knowledge, ensuring accurate interpretation and operation of downstream tasks at the destination.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure KR2023006128_07112024_PF_FP_ABST
    Figure KR2023006128_07112024_PF_FP_ABST
Patent Text Reader

Abstract

A method performed by a first node in a wireless communication system, according to an embodiment of the present disclosure, may comprise the steps of: receiving at least one synchronization signal from a second node; receiving control information from the second node; receiving information related to semantic communication from the second node; on the basis of the information related to the semantic communication and background knowledge, extracting at least one piece of raw data from each of a plurality of data modalities; performing encoding on the at least one piece of raw data; acquiring at least one semantic representation vector set on the basis of the at least one piece of raw data on which the encoding has been performed; generating at least one semantic representation packet on the basis of the at least one semantic representation vector set; transmitting the at least one semantic representation packet to the second node; receiving update information from the second node; and updating the background knowledge on the basis of the update information.
Need to check novelty before this filing date? Find Prior Art

Citation Information

Patent Citations

  • KR20210154705A