About this document

Generalization - Principle (2) TAG, Node Etc by Atharva Pratap Singh is a document available to read on EtoBox.

This document presents a framework called LLM-BP for improving inference over text-attributed graphs (TAGs) using large language models (LLMs). It establishes two key principles: unifying the text space with task-adaptive embeddings and developing a generalizable graph information aggregation mechanism, which significantly enhances zero-shot generalization performance. Evaluations on 11 real-world TAG benchmarks show that LLM-BP outperforms existing methods, achieving notable improvements in node labeling t

Author

Atharva Pratap Singh

Language

EN