Order Matters in Retrosynthesis: Structure-aware Generation via Reaction-Center-Guided Discrete Flow Matching
Chenguang Wang, Zihan Zhou, Lei Bai, Tianshu Yu · 2026
Read Order Matters in Retrosynthesis: Structure-aware Generation via Reaction-Center-Guided Discrete Flow Matching by Chenguang Wang, Zihan Zhou, Lei Bai, Tianshu Yu (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-2602-13136.txt.
Description
Retrosynthesis asks which starting materials could yield a target molecule, and template-free methods treat it as opaque sequence generation. Chenguang Wang and three co-authors argue that atom order in neural representations matters, and build a structure-aware framework around it. By placing reaction-centre atoms at the head of the sequence, the method turns implicit chemical knowledge into positional patterns a model can grasp.
The RetroDiT backbone, a graph transformer using rotary position embeddings, uses that ordering to focus on chemically critical regions. Paired with discrete flow matching, training and sampling separate, and generation takes 20 to 50 steps versus 500 for prior diffusion methods. Results reach 61.2% top-1 on USPTO-50k and 51.3% on USPTO-Full with predicted centres, rising to 71.1% and 63.4% with oracle centres, while a tiny 280K-parameter ordered model matches a 65M-parameter unordered one.
Catalog record
| CREATOR | Chenguang Wang, Zihan Zhou, Lei Bai, Tianshu Yu |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 98 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2602-13136