The linear autoencoder extracts the dimension reduced feature vector zi from the flattened feature vector ̂xi as z i = W ̂x i + b (2) where W ∈ Rd× k and b ...
確定! 回上一頁