USPTO-50k
PulseAugur coverage of USPTO-50k — every cluster mentioning USPTO-50k across labs, papers, and developer communities, ranked by signal.
-
New LLM RetroDFM-R offers reasoning-driven retrosynthesis prediction
Researchers have developed RetroDFM-R, a new large language model designed for retrosynthetic planning in organic synthesis and drug discovery. This model utilizes reinforcement learning to provide not just accurate pre…
-
LLMs advance chemical retrosynthesis prediction with new models and frameworks · 4 sources tracked
Researchers are developing advanced large language models (LLMs) for chemical retrosynthesis, aiming to improve the prediction of synthetic routes for molecules. One approach, C3LM, utilizes a massive dataset of ~45.6 m…
-
New RetroDiT framework improves chemical reaction prediction with structure-aware AI
Researchers have developed a new structure-aware framework for template-free retrosynthesis, a process crucial for chemical reaction prediction. This method, called RetroDiT, utilizes a graph transformer with rotary pos…
-
RETROSPECT system boosts chemical retrosynthesis prediction accuracy
Researchers have developed RETROSPECT, a novel system for chemical retrosynthesis that improves prediction accuracy and candidate selection. The system utilizes a Transformer-based proposal model, ChemAlign, combined wi…
-
New GRG method boosts AI retrosynthesis accuracy and speed
Researchers have developed a new method called Graph-oriented Representation Guidance (GRG) to improve molecular graph generators for retrosynthesis. This technique guides the generator with representations from pre-tra…