ChemFusion neural network predicts reaction yields with novel multimodal approach

Researchers have developed ChemFusion, a novel multimodal neural network designed to predict the outcomes of transition-metal-catalyzed reactions. This hybrid network effectively merges electronic descriptors with the 3D geometry of reactive sites using a cross-attention mechanism. By dynamically linking global electronic states to specific spatial constraints, ChemFusion demonstrates superior predictive performance compared to single-modality frameworks and offers physically grounded interpretability by identifying steric hindrances. AI

IMPACT This model could accelerate chemical research by improving the accuracy and interpretability of reaction outcome predictions.

RANK_REASON The cluster contains a research paper detailing a new model for reaction yield prediction. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.LG →

• alphaXiv
• arXiv
• CatalyzeX Code Finder for Papers
• ChemFusion
• CORE Recommender
• Cross-Couplings Using Aryl Ethers via C–O Bond Activation Enabled by Nickel Catalysts
• DagsHub
• Gotit.pub
• Hugging Face
• IArxiv Recommender
• ScienceCast

• paper
• model release

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →