Summary of Large Generative Graph Models, by Yu Wang et al.
Large Generative Graph Models
by Yu Wang, Ryan A. Rossi, Namyong Park, Huiyuan Chen, Nesreen K. Ahmed, Puja Trivedi, Franck Dernoncourt, Danai Koutra, Tyler Derr
First submitted to arxiv on: 7 Jun 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 The abstract proposes a new class of graph generative models called Large Graph Generative Models (LGGM) that are trained on a large corpus of graphs from 13 different domains. The authors demonstrate the pre-trained LGGM has superior zero-shot generative capability to existing graph generative models and can be fine-tuned for specific tasks. Additionally, they equip the LGGM with Text-to-Graph capability, allowing users to generate graphs given text prompts. The proposed model is designed to integrate extensive world knowledge from language models, providing fine-grained control of generated graphs. |
Low | GrooveSquid.com (original content) | Low Difficulty Summary Large graph generative models (LGMs) like GPT and Stable Diffusion are trained on a vast amount of data. But previous graph generative models were only trained on one dataset at a time. To fix this, the authors created LGGM, a new class of graph generative model that’s trained on many different graphs from 13 domains. They show that LGGM is better than other graph generative models at generating new graphs without any extra training. They also make it possible to fine-tune LGGM for specific tasks and even generate graphs based on text prompts. |
Keywords
» Artificial intelligence » Diffusion » Generative model » Gpt » Zero shot