Very usefulr article by NetFlix: http://techblog.netflix.com/2014/02/distributed-neural-networks-with-gpus.html

• spearmint: Hyperparameter tuning by Bayesian optimization
• Celery: distributed task queue library.

Completed: Multilayer perception tutorial with Theano.

• Feedforward artificial neural network model
• "Logistic regression is a special case of the MLP with no hidden layer (the input is directly connected to the output) and the cross-entropy (sigmoid output) or negative log-likelihood (softmax output) loss." from here.
• Cross-entropy loss L(f, (x,y)) is minimized when x=y.

To Read:

• Understanding the difficulty of training deep feedforward neural networks
• Practical Bayesian Optimization of Machine Learning Algorithms (re-read)
• Efficient BackProp