1 paper · 1 filter
Akash Srivastava, Yamini Bansal, Yukun Ding +15
Current autoencoder-based disentangled representation learning methods achieve disentanglement by penalizing the (aggregate) posterior to encourage statistical independence of the…