Federated Learning from Molecules to Processes: A Perspective
Jan G. Rittig, Clemens Kortmann · 2025
Read Federated Learning from Molecules to Processes: A Perspective by Jan G. Rittig, Clemens Kortmann (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-2506-18525.txt.
Description
Jan Rittig and Clemens Kortmann argue that chemical companies should train shared machine learning models without ever pooling their data. Process and molecular data are trade secrets, locked in company silos, which starves models of scale. Federated learning trains one model across many owners while raw data stays home. Their 2025 perspective maps where the trick fits chemical engineering, from molecules to full plants.
Two simulated collaborations test the idea: predicting mixture behavior with graph networks, and identifying a distillation column's dynamics with autoencoders. Jointly trained models beat what each company could build alone. Industry modelers get both a map of applications and proof the mechanics work on chemical problems.
Catalog record
| CREATOR | Jan G. Rittig, Clemens Kortmann |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 62 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2506-18525