TauFlowNet: Revealing latent propagation mechanism of tau aggregates using deep neural transport equations

Tingting Dan, Mustafa Dere, Won Hwa Kim, Minjeong Kim, Guorong Wu

Abstract: Highlights•We build a bridge among graph neural networks, partial differential equations and calculus of variations.•We introduce the total variation (TV) into the graph transport model for maximizing the spreading flow while minimizing the overall potential energy.•We design a generative adversarial network (GAN) to characterize the TV-based spreading flow of tau aggregates, coined TauFlowNet.