CatFlow : Co-generation of Slab-Adsorbate Systems via Flow Matching
Minkyu Kim, Nayoung Kim, Honghui Kim, Sungsoo Ahn · 2026
Read CatFlow : Co-generation of Slab-Adsorbate Systems via Flow Matching by Minkyu Kim, Nayoung Kim, Honghui Kim, Sungsoo Ahn (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-2602-05372.txt.
Description
Designing a catalyst surface apart from the molecule that sits on it misses how the two shape each other. Minkyu Kim and 3 co-authors present CatFlow in this 2026 materials science paper, an AI scheme that creates slab and adsorbate at once. Flow matching, a modern AI technique, drives the co-generation.
The setup is the key economy. The slab-adsorbate complex is described on a primitive cell in factored form, cutting the count of learnable knobs by about nine times on average. It plainly encodes the surface tilt of the join. Trained and tested on the Open Catalyst 2020 dataset, CatFlow produced shapes with better match than step-by-step and sequential baselines. Made joins also matched realistic adsorption energy spreads and sat closer to energy local minima, which is what a chemist screening options truly needs from a generator.
Catalog record
| CREATOR | Minkyu Kim, Nayoung Kim, Honghui Kim, Sungsoo Ahn |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 50 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2602-05372