Fig. 1
From: A deep auto-encoder model for gene expression prediction
An Illustration of an Auto-encoder Corruption Model. The raw input X is corrupted via process q. The black nodes denote the corrupted input. The corrupted input is converted to Y via process f θ . Afterwards, Y attempts to reconstruct the raw input via process g θ , and generates the reconstruction Z. A loss function L(X,Z) is used to calculate the reconstruction error for backpropagation