From here you can access the Problem Sets when they become available:

• Problem Set 1. Integrate-and-fire neurons (Postscript, PDF). Due Sept 10.
• Problem Set 2. Simple perceptrons and gradient descent (Postscript, PDF). Due Sept 17.
  • Data file: mnistabridged.mat or mnistabridged.zip (for Windows users that have trouble downloading the .mat properly).
  • Code Sample: two.m
• Problem Set 3. Multilayer perceptrons and backpropagation (Postscript, PDF). Due Sept 24.
  • Data file: mnistabridged.mat
• Problem Set 4. Statistical learning theory (Postscript, PDF). Due Oct. 1.
• Problem Set 5. PCA and linear networks (Postscript, PDF). Due Oct. 8.
• Problem Set 6. Nonlinear network theory (Postscript, PDF). Due Oct. 22.
• Problem Set 7. Nonlinear network theory again (Postscript, PDF). Due Oct. 31.
• Problem Set 8. VQ and Lyapunov (Postscript, PDF). Due Nov. 12.
• Problem Set 9. Models of associative memory (Postscript, PDF). Due Nov. 19.
• Problem Set 10. Neural integrators and contrastive Hebbian Learning (Postscript, PDF). Due Nov. 26.
• Problem Set 11. Reinforcement and trajectory learning (Postscript, PDF). Due Dec. 5.