How does topology of neural architectures impact gradient propagation and model performance?

1 code implementation • CVPR 2021 • Kartikeya Bhardwa, Guihong Li2, Radu Marculescu

In this paper, we reveal that the topology of the concatenation-type skip connections is closely related to the gradient propagation which, in turn, enables a predictable behavior of DNNs’ test performance.

Model Compression Neural Architecture Search