Researchers have developed Embedded Graph Flows (EGF), a novel generative model designed for categorical graph generation. Unlike previous methods that use fixed one-hot vectors for categories, EGF learns continuous embeddings for node and edge types, transporting Gaussian noise through a permutation-equivariant graph transformer. This approach allows for more flexible and accurate representation of categorical data. EGF has demonstrated competitive performance on molecular benchmarks, achieving state-of-the-art results on the QM9 dataset and showing strong agreement with reference molecules on the ZINC250k dataset. AI
IMPACT This new model could improve the accuracy and flexibility of generating complex categorical data, particularly in molecular discovery and drug development.
RANK_REASON The cluster contains a research paper detailing a new model and benchmark results. [lever_c_demoted from research: ic=1 ai=1.0]
- Embedded Graph Flows
- Maximum Mean Discrepancy
- neighbourhood subgraph pairwise distance kernel
- NSPDK
- QM9
- ZINC250k
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →