Fig. 1From: On tower and checkerboard neural network architectures for gene expression inferenceThe original C–D–GEX architectures and the novel architectures proposed in this paper. The outputs of the towers (or halves for C–D–GEX) are averaged before the output layerBack to article page