Molecular property prediction has long been a critical challenge in drug development. Traditional quantum chemistry methods are inefficient when dealing with large-scale molecular databases, failing ...
Classic Graph Convolutional Networks (GCNs) often learn node representation holistically, which would ignore the distinct impacts from different neighbors when aggregating their features to update a ...
Some results have been hidden because they may be inaccessible to you
Show inaccessible results
Feedback