Accelerating the inference of string generation-based chemical reaction models for industrial applications
Uloženo v:
| Název: | Accelerating the inference of string generation-based chemical reaction models for industrial applications |
|---|---|
| Autoři: | Mikhail Andronov, Natalia Andronova, Michael Wand, Jürgen Schmidhuber, Djork-Arné Clevert |
| Zdroj: | Journal of Cheminformatics, Vol 17, Iss 1, Pp 1-11 (2025) |
| Informace o vydavateli: | BMC, 2025. |
| Rok vydání: | 2025 |
| Sbírka: | LCC:Information technology LCC:Chemistry |
| Témata: | CASP, Speculative decoding, Single-step retrosynthesis, Fast inference, Reaction prediction, Information technology, T58.5-58.64, Chemistry, QD1-999 |
| Popis: | Abstract Transformer-based, template-free SMILES-to-SMILES translation models for reaction prediction and single-step retrosynthesis are of interest to computer-aided synthesis planning systems, as they offer state-of-the-art accuracy. However, their slow inference speed limits their practical utility in such applications. To address this challenge, we propose speculative decoding with a simple chemically specific drafting strategy and apply it to the Molecular Transformer, an encoder-decoder transformer for conditional SMILES generation. Our approach achieves over 3X faster inference in reaction product prediction and single-step retrosynthesis with no loss in accuracy, increasing the potential of the transformer as the backbone of synthesis planning systems. To accelerate the simultaneous generation of multiple precursor SMILES for a given query SMILES in single-step retrosynthesis, we introduce Speculative Beam Search, a novel algorithm tackling the challenge of beam search acceleration with speculative decoding. Our methods aim to improve transformer-based models’ scalability and industrial applicability in synthesis planning. |
| Druh dokumentu: | article |
| Popis souboru: | electronic resource |
| Jazyk: | English |
| ISSN: | 1758-2946 |
| Relation: | https://doaj.org/toc/1758-2946 |
| DOI: | 10.1186/s13321-025-00974-w |
| Přístupová URL adresa: | https://doaj.org/article/198136afc55f4deb97bf360c11208e10 |
| Přístupové číslo: | edsdoj.198136afc55f4deb97bf360c11208e10 |
| Databáze: | Directory of Open Access Journals |
Full Text 
Full Text Finder 
Nájsť tento článok vo Web of Science | Abstrakt: | Abstract Transformer-based, template-free SMILES-to-SMILES translation models for reaction prediction and single-step retrosynthesis are of interest to computer-aided synthesis planning systems, as they offer state-of-the-art accuracy. However, their slow inference speed limits their practical utility in such applications. To address this challenge, we propose speculative decoding with a simple chemically specific drafting strategy and apply it to the Molecular Transformer, an encoder-decoder transformer for conditional SMILES generation. Our approach achieves over 3X faster inference in reaction product prediction and single-step retrosynthesis with no loss in accuracy, increasing the potential of the transformer as the backbone of synthesis planning systems. To accelerate the simultaneous generation of multiple precursor SMILES for a given query SMILES in single-step retrosynthesis, we introduce Speculative Beam Search, a novel algorithm tackling the challenge of beam search acceleration with speculative decoding. Our methods aim to improve transformer-based models’ scalability and industrial applicability in synthesis planning. |
|---|---|
| ISSN: | 17582946 |
| DOI: | 10.1186/s13321-025-00974-w |