Scaling equilibrium propagation to deep convnets by drastically reducing its gradient estimator bias

HIGHLIGHTS

• who: Axel Laborieux and Maxence Ernoult from the Université Paris-Saclay, CNRS, Centre de Nanosciences et de Nanotechnologies, Palaiseau, France, Unité Mixte de Heidelberg University, Germany have published the Article: Scaling Equilibrium Propagation to Deep ConvNets by Drastically Reducing Its Gradient Estimator Bias, in the Journal: (JOURNAL)
• what: The authors show that performing the second phase of EP with nudging strength of constant sign induces a systematic first order bias in the EP gradient estimate which, once canceled, unlocks the training of deep convolutional neural_networks (ConvNets), with bidirectional or unidirectional connections and with performance closely . . .
  

   

  

  If you want to have access to all the content you need to log in!

  Thanks :)

  Username or Email

  Password

  Remember me

  Lost your password?

  If you don't have an account, you can create one here.