LLMs and LVLMs-Driven Methods for Graph Processing

PhD Qualifying Examination


Title: "LLMs and LVLMs-Driven Methods for Graph Processing"

by

Mr. Yanbin WEI


Abstract:

Large Language Models (LLMs) and Large Vision-Language Models (LVLMs) have
showcased exceptional problem-solving capabilities across various domains,
sparking interest among graph researchers to explore their potential in graph-
related contexts, such as academic and e-commerce networks. However, LLMs and
LVLMs often face challenges with graph-related tasks due to the mismatch between
their training data—primarily natural language and images, and the structural
nature of graph topologies. To overcome this challenge, researchers have
developed methods to enhance these models' proficiency in processing graphs,
thereby boosting their effectiveness in graph-related tasks. This work presents a
comprehensive review of these advancements. Specifically, we systematically
categorize the existing methods based on their strategies for processing graph
information: i) Transforming graphs into textual descriptions; ii) Encoding
graphs as numerical embeddings; iii) Depicting graphs through visual formats;
and iv) Integrating external tools for graph data retrieval and processing.
Notably, this survey distinguishes itself from previous "LLMs for Graphs"
reviews by focusing exclusively on studies where LLMs/LVLMs serve as the
primary problem solver, termed LLM/LVLM-driven graph methods, and excluding
those where they are merely used as auxiliary feature-processing modules. This
distinction is crucial for clarity. In addition to the taxonomy, the survey
examines key applications, practical resources, and highlights open challenges
and future research directions.


Date:                   Thursday, 26 February 2026

Time:                   10:00am - 11:30am

Venue:                  Room 2132C
                        Lift 22

Committee Members:      Prof. James Kwok (Supervisor)
                        Prof. Raymond Wong (Chairperson)
                        Dr. Yangqiu Song