Deep learning for Chemometric and non-translational data

Jacob Søgaard Larsen, Line Clemmensen

We propose a novel method to train deep convolutional neural networks which learn from multiple data sets of varying input sizes through weight sharing. This is an advantage in che…