An Alternative to Restricted-Boltzmann Learning for Binary Latent Variables based on the Criterion of Maximal Mutual Information.

David Edelman

Browse the full ICAART paper archive.