Papers › Molecular Graph Enhanced Transformer for Retrosynthesis Prediction
Molecular Graph Enhanced Transformer for Retrosynthesis Prediction
Kelong Mao, Peilin Zhao, Tingyang Xu, Yu Rong, Xi Xiao, Junzhou Huang
With massive possible synthetic routes in chemistry, retrosynthesis prediction is still a challenge for researchers. Recently, retrosynthesis prediction is formulated as a Machine Translation (MT) task. Namely, since each molecule can be represented as a Simplified Molecular-Input Line-Entry System (SMILES) string, the process of synthesis is analogized to a process of language translation from reactants to products. However, the MT models that applied on SMILES data usually ignore the information of natural atomic connections and the topology of molecules. In this paper, we propose a Graph Enhanced Transformer (GET) framework, which adopts both the sequential and graphical information of molecules. Four different GET designs are proposed, which fuse the SMILES representations with atom embedding learned from our improved Graph Neural Network (GNN). Empirical results show that our model significantly outperforms the Transformer model in test accuracy.
Code
No code repository is listed for this paper in the archive or in Syntology's graph.
Code Syntology ran Syntology
Not run by Syntology. Nothing on this page verifies that the listed code works.
Tasks
Results from the paper archive 2025-07-28
| Task | Dataset | Model | Metric | Value | Rank at snapshot | Leaderboard | Report |
|---|---|---|---|---|---|---|---|
| Single-step retrosynthesis | USPTO-50k | GET-LT1 (reaction class as prior) | Top-1 accuracy | 57.4 | #11 of 35 | Archive leaderboard | report |
| Single-step retrosynthesis | USPTO-50k | GET-LT1 (reaction class as prior) | Top-10 accuracy | 77.4 | #11 of 35 | Archive leaderboard | report |
| Single-step retrosynthesis | USPTO-50k | GET-LT1 (reaction class as prior) | Top-3 accuracy | 71.3 | #11 of 35 | Archive leaderboard | report |
| Single-step retrosynthesis | USPTO-50k | GET-LT1 (reaction class as prior) | Top-5 accuracy | 74.8 | #11 of 35 | Archive leaderboard | report |
| Single-step retrosynthesis | USPTO-50k | GET-LT2 (reaction class as prior) | Top-1 accuracy | 56.2 | #12 of 35 | Archive leaderboard | report |
| Single-step retrosynthesis | USPTO-50k | GET-LT2 (reaction class as prior) | Top-10 accuracy | 74.7 | #12 of 35 | Archive leaderboard | report |
| Single-step retrosynthesis | USPTO-50k | GET-LT2 (reaction class as prior) | Top-3 accuracy | 69.4 | #12 of 35 | Archive leaderboard | report |
| Single-step retrosynthesis | USPTO-50k | GET-LT2 (reaction class as prior) | Top-5 accuracy | 72.5 | #12 of 35 | Archive leaderboard | report |
| Single-step retrosynthesis | USPTO-50k | GET-LT1 (reaction class unknown) | Top-1 accuracy | 44.9 | #33 of 35 | Archive leaderboard | report |
| Single-step retrosynthesis | USPTO-50k | GET-LT1 (reaction class unknown) | Top-10 accuracy | 65.9 | #33 of 35 | Archive leaderboard | report |
| Single-step retrosynthesis | USPTO-50k | GET-LT1 (reaction class unknown) | Top-3 accuracy | 58.8 | #33 of 35 | Archive leaderboard | report |
| Single-step retrosynthesis | USPTO-50k | GET-LT1 (reaction class unknown) | Top-5 accuracy | 62.4 | #33 of 35 | Archive leaderboard | report |
Ranks are positions in the archive's leaderboards as they stood at the 2025-07-28 snapshot. Results published since then are not among these rows, so a rank here is not a current standing.
Methods
Report a problem or propose a change · a person checks every report against the paper or source before anything changes; decisions are listed on /corrections