ItDPDM: Information-Theoretic Discrete Poisson Diffusion Model
Sagnik Bhattacharya, Abhiram R. Gorle, Ahsan Bilal, Connor Ding, Amit Kumar Singh Yadav, Tsachy Weissman · 2025
Read ItDPDM: Information-Theoretic Discrete Poisson Diffusion Model by Sagnik Bhattacharya, Abhiram R. Gorle, Ahsan Bilal, Connor Ding, Amit Kumar Singh Yadav, Tsachy Weissman (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-2505-05082.txt.
Description
Six authors led by Sagnik Bhattacharya build a new kind of AI model for data that comes in whole counts. Such data cover things like music notes, where values are never parts or below zero. Many past models first turn counts into smooth values, which fits the data poorly. Others train on a loose bound, so their odds scores for new data are off and samples suffer.
The new model, called ItDPDM, works in the count space the whole way through. It draws on the physics of light hits, which also arrive as counts. A new loss term has a proven exact link to the true odds of the data. Tests on made-up count sets, music, and images show better odds scores than older models. Sample quality held up well next to the best past work.
Catalog record
| CREATOR | Sagnik Bhattacharya, Abhiram R. Gorle, Ahsan Bilal, Connor Ding, Amit Kumar Singh Yadav, Tsachy Weissman |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 168 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2505-05082