Is Noise Conditioning Necessary? A Unified Theory of Unconditional Graph Diffusion Models
Jipeng Li, Yanning Shen · 2025
Read Is Noise Conditioning Necessary? A Unified Theory of Unconditional Graph Diffusion Models by Jipeng Li, Yanning Shen (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2505-22935.txt.
Introduction
Jipeng Li and Yanning Shen ask whether graph diffusion models really need to be told the noise level. Conditioning on noise is standard practice, yet corrupted graphs may carry the clue themselves. The team builds theory around Bernoulli edge-flip corruption and extends it to coupled noise on structure and attributes. Tests ran on synthetic and real graphs with GDSS and DiGress models.
Unconditional models matched or beat their conditioned twins. They also used 4 to 6% fewer parameters and ran 8 to 10% faster. The authors argue that high-dimensional graph data often encodes enough noise information on its own, opening simpler designs.
Catalog record
| CREATOR | Jipeng Li, Yanning Shen |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 140 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2505-22935