- (2024, WIREs Comput Mol Sci) Recent advances in deep learning for retrosynthesis. [Page]
- (Syntheseus, Faraday Discuss) Re-evaluating Retrosynthesis Algorithms with Syntheseus. [Link]
- (Digital Discovery) Models matter: the impact of single-step retrosynthesis on synthesis planning. [Link]
- (EditRetro, Nat Commun) Retrosynthesis prediction with an iterative string editing model. [Page]
- (Digital Discovery) Investigating the reliability and interpretability of machine learning frameworks for chemical retrosynthesis. [Page]
- (ReSynZ, J Chem Theory Comput) Retrosynthesis Zero: self-improving global synthesis planning using reinforcement learning.
[Page]
- Policy improvement with MCTS suggested actions.
- (MEEA*, Commun Chem) Efficient retrosynthetic planning with MCTS exploration enhanced A* search.
[Page]
- A* search with exploration from MCTS.
- (Graph2Edit, Nat Commun) Retrosynthesis prediction using an end-to-end graph generative architecture for molecular graph editing.
[Page]
- Semi-template retrosynthesis with edit vocabulary and end-to-end architecture.
- (R-SMILES, Chem Sci) Root-aligned SMILES: a tight representation for chemical reaction prediction.
[Page]
- Molecule representation with minimal edit distance between product and reactants.
- (GraphRetro, NeurIPS) Learning graph models for retrosynthesis prediction.
[Page]
- Semi-template retrosynthesis with subgraph vocabulary.
- (Kim, ICML) Self-improving retrosynthesis planning.
[Page]
- Backward reaction model learned by imitating RETRO*-generated realistic data.
- (Retro*, ICML) Retro*: learning retrosynthetic planning with neural guided A* search.
[Page]
- Pioneer work in retrosynthetic planning; search with a value function.
- (Schreck, ACS Cent Sci) Learning retrosynthetic planning through simulated experience.
[Page]
- Retrosynthesis with deep reinforcement learning.
- USPTO-50k [Page]
- USPTO-MIT
- USPTO-FULL