Search papers, labs, and topics across Lattice.
Duke University, Georgia Institute of Technology
2
0
2
ChemFusion's cross-attention mechanism reveals that spatially aware networks can outperform traditional models by effectively navigating complex steric constraints in chemical reactions.
Integrating 3D geometry, topological grammar, and physicochemical descriptors leads to a 20.6% reduction in prediction error for molecular properties, all while using under one million parameters.
