Summary of Zerog: Investigating Cross-dataset Zero-shot Transferability in Graphs, by Yuhan Li et al.
ZeroG: Investigating Cross-dataset Zero-shot Transferability in Graphs
by Yuhan Li, Peisong Wang, Zhixun Li, Jeffrey Xu Yu, Jia Li
First submitted to arxiv on: 17 Feb 2024
Categories
- Main: Machine Learning (cs.LG)
- Secondary: None
GrooveSquid.com Paper Summaries
GrooveSquid.com’s goal is to make artificial intelligence research accessible by summarizing AI papers in simpler terms. Each summary below covers the same AI paper, written at different levels of difficulty. The medium difficulty and low difficulty versions are original summaries written by GrooveSquid.com, while the high difficulty version is the paper’s original abstract. Feel free to learn from the version that suits you best!
| Summary difficulty | Written by | Summary |
|---|---|---|
| High | Paper authors | High Difficulty Summary Read the original abstract here |
| Medium | GrooveSquid.com (original content) | Medium Difficulty Summary In this study, researchers introduce ZeroG, a new framework that enables zero-shot transferability in graphs. The approach addresses challenges like feature misalignment and mismatched label spaces by leveraging a language model to encode node attributes and class semantics. A prompt-based subgraph sampling module is also proposed to enrich semantic information and structure information. A lightweight fine-tuning strategy reduces the risk of overfitting while maintaining zero-shot learning efficacy. The results demonstrate significant cross-dataset zero-shot transferability, paving the way for graph foundation models. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This study creates a new framework called ZeroG that lets graphs learn without needing data labels. It solves problems like features not matching and different label meanings by using a language model to understand node attributes and class meanings. The study also proposes a way to sample subgraphs based on prompts, which helps with learning. A small fine-tuning step is used to prevent the model from getting too good at remembering specific data. The results show that ZeroG can learn from one graph without seeing others, opening up new possibilities for creating foundation models. |
Keywords
* Artificial intelligence * Fine tuning * Language model * Overfitting * Prompt * Semantics * Transferability * Zero shot
