Training Very Deep Networks — mentioned by jisaacso

Tue, 26 Jul 2016 14:30:51 GMT

Hey thanks for the great questions! (1) The NN uses two convolutional units and a fully connected softmax layer. Relative to Inception V3 or highway networks, this is _not_ a very deep architecture. I was looking for a balance between accuracy and training time (trained on my MBP). (2) I looked into other neural architectures (LSTM and several fully connected) without much difference in performance. If you take a look at the physionet google group, there are a number of other methods evaluated (logistic regression, SVMs, etc.) (3) I did vary hyperparameters and saw a dropout of ~0.45 and…

Paper and discussion history. A mention is not an endorsement.

Training Very Deep Networks · HackerPapers discussions

Training Very Deep Networks — mentioned by jisaacso