Talk like a Graph: Encoding Graphs for Large Language Models
This addresses the understudied challenge of enabling LLMs to reason on graphs for applications like social networks and recommender systems, representing an incremental advance in graph encoding techniques.
The paper tackled the problem of encoding graph-structured data as text for large language models (LLMs) to improve graph reasoning, showing that performance varies based on encoding methods, graph tasks, and graph structures, with correct encoder choices boosting performance by 4.8% to 61.8% on specific tasks.
Graphs are a powerful tool for representing and analyzing complex relationships in real-world applications such as social networks, recommender systems, and computational finance. Reasoning on graphs is essential for drawing inferences about the relationships between entities in a complex system, and to identify hidden patterns and trends. Despite the remarkable progress in automated reasoning with natural text, reasoning on graphs with large language models (LLMs) remains an understudied problem. In this work, we perform the first comprehensive study of encoding graph-structured data as text for consumption by LLMs. We show that LLM performance on graph reasoning tasks varies on three fundamental levels: (1) the graph encoding method, (2) the nature of the graph task itself, and (3) interestingly, the very structure of the graph considered. These novel results provide valuable insight on strategies for encoding graphs as text. Using these insights we illustrate how the correct choice of encoders can boost performance on graph reasoning tasks inside LLMs by 4.8% to 61.8%, depending on the task.