A learning algorithm for cortex
How does the brain learn its internal representations?
The standard way to find the serious rivals
Autoencoders as generative models
Bottom-up and top-down weights can be learned without considering the partition function