How Temporal Correlations Shape Memory in Linear Recurrent Neural Networks
One property of the input is turned into a prediction of whether a network learns memory and why correlated data turns recurrent networks into change detectors and explains why correlated data turns recurrent networks into change detectors.
Arnol Fokam, Fasseu Sieyondji Akpevwoghene, Edem Fiifi Dawson
· 0 citations