A Multimodal Conditional Mixture Model with Distribution-Level Physics Priors
Jinkyo Han, Bahador Bahmani · 2026
Read A Multimodal Conditional Mixture Model with Distribution-Level Physics Priors by Jinkyo Han, Bahador Bahmani (2026) 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-2602-10451.txt.
Description
Jinkyo Han and Bahador Bahmani model systems that can behave in several distinct ways at once in a 2026 paper. Many real and building systems switch between regimes, so their outcomes form multiple modes rather than one tidy spread. Standard machine learning captures the modes but can violate physics; physics models stay consistent but struggle with limited data.
Their system uses mixture density networks, which represent a full conditional spread as a mix of parts, and embeds real knowledge through part-exact regularization. Each part can carry its own real constraints, keeping the result clear. The approach suits settings where data are scarce and real consistency matters, from building design to scientific prediction under regime switching.
Catalog record
| CREATOR | Jinkyo Han, Bahador Bahmani |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 44 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2602-10451