Reverse-engineering deep neural networks | Ilya Kuprov

Speaker: Prof Ilya Kuprov, Professor of physics at the University of Southampton

The lack of interpretability is a much-criticised feature of deep neural networks. Often, a neural network is effectively a black box. However, we have recently found a group-theoretical procedure that brings inner layer signalling into a human-readable form. We applied it to a signal processing network used in magnetic resonance spectroscopy, and found that the network spontaneously invents a bandpass filter, a notch filter, a frequency axis rescaling transformation, frequency division multiplexing, group embedding, spectral filtering regularisation, and a map from harmonic functions into Chebyshev polynomials – in ten minutes of unattended training.

This talk is part of the Liverpool Virtual Seminar Series on Data Intensive Science; more information can be found at https://indico.ph.liv.ac.uk/e/data_sc...

#bigdata #datascience #dataseminar #science #data #spectroscopy