Skip to main content
. 2020 Aug 11;11(35):9524–9531. doi: 10.1039/d0sc02458a

Fig. 1. Overview of the method. Adjacency matrix E is omitted from the figure. FC stands for fully-connected layer and MLP stands for multilayer perceptron. Concat denotes concatenation. NNConv and GRU are explained in eqn (1) and (2), respectively. “Normalize” means L2 normalization.

Fig. 1