Filling in the Gap: a General Method Using neural Networks

When a set of medical signals has redundant informa-tion, it is sometimes possible to recover one signal, fromits past and the information provided by the other signals.In this work, we present a general method to realize thattask. It has been known for a long time that multilayerednetworks are universal approximators, but, even with thebackprop algorithm, it was not possible to train such a net-work, to realize complex real life tasks. In the last years,Geoffrey Hinton presented a training strategy that allowsto overcome the previous difficulties. We describe a way ofadapting Hinton’s strategy to our task. An example of a situation considered here, consists ontraining a Multilayered perceptron to take ECG leads IIand I as input and produce as output missing lead V. This method got the best scores among participants inthe Physionet/ Computing in Cardiology Challenge 2010.