Advancing semantic communication systems through knowledge graphs, generative AI, and safeguarded AI

Liang, Chengsi (2026) Advancing semantic communication systems through knowledge graphs, generative AI, and safeguarded AI. PhD thesis, University of Glasgow.

Abstract

Semantic communication (SemCom) represents a paradigm shift that prioritizes conveying information meaning over traditional bit-level transmission. However, SemCom faces fundamental challenges in explicitly and logically characterizing semantics within coding models. Knowledge graphs (KGs) emerge as a promising solution by encapsulating entity attributes and relational logic through structured triples of entities and relationships. The multi-modal nature of KGs, encompassing text, images, and audio data, enables comprehensive semantic representation across diverse communication scenarios.
Despite their potential, integrating KGs into SemCom systems presents three critical challenges. First, developing effective methods to align and integrate source data with KG information for coherent semantic representations remains complex. Second, reconstructing original data from KGs proves particularly difficult under adverse communication conditions. Third, the KG integration inevitably introduces additional transmission overhead that must be carefully managed.
This thesis addresses these challenges by designing and optimizing KG-based SemCom frameworks across multiple data formats. The research comprises five interconnected contributions. The first work establishes a KG-based SemCom framework for video delivery that provides foundational principles for subsequent VR applications. The second work investigates generative AI-driven SemCom networks incorporating KG utilization. The third work develops a KG-based SemCom framework for audio delivery in Internet of Sounds environments. The fourth work presents a comprehensive KG-enabled SemCom framework with detailed KG fusion methodology. The final work addresses AI safety concerns by proposing a safeguarded AI SemCom framework for secure SemCom systems.
In conclusion, the work presented in this thesis provides insight into the design of KGempowered SemCom systems for multi-modal data transmission, which can be viewed as a foundational step towards achieving efficient semantic representation and reconstruction.

• Item Type: Thesis (PhD)
• Qualification Level: Doctoral
• Subjects: T Technology > TA Engineering (General). Civil engineering (General)
• Colleges/Schools: College of Science and Engineering > School of Engineering
• Supervisor's Name: Sun, Dr. Yao and Imran, Professor Muhammad
• Date of Award: 2026
• Depositing User: Theses Team
• Unique ID: glathesis:2026-86070
• Copyright: Copyright of this thesis is held by the author.
• Date Deposited: 24 Jun 2026 15:19
• Last Modified: 24 Jun 2026 15:23
• Thesis DOI: 10.5525/gla.thesis.86070
• URI: https://theses.gla.ac.uk/id/eprint/86070

Related URLs

• Article DOI
• Article DOI
• Article DOI
• Article DOI
• Article DOI
• Article DOI
• Conference proceeding
• Conference proceeding