Skip to main content
. 2021 Nov 19;29(1):72–79. doi: 10.1093/jamia/ocab229

Figure 4.

Figure 4.

The structure of 3 deep learning models: (A), (B), and (C) are the structures of NN Baseline, SLAT, and MT-SLAT, respectively. Color coding is consistent with Figure  1. NN block denotes the neural network structure with multiple layers. Bottleneck block denotes a one-layer neural network whose dimension is smaller than other NN blocks. The setting for each block is provided in the Experimental Setting section. NN: neural network; SLAT: Shared LATent layer;