EEG P300 data generation demo and classification

[toncho11]

A joint p-bit probabilistic model for P300 EEG that does:

• primary: stochastic conditional EEG generation with a p-bit PCircuit.
• secondary: the same circuit also gives competitive classification using exact conditional classification

The demo runs in 3 modes:

• classify - compares classical FlatLR with JointPBit, a hybrid Fourier + PCircuit version (of FlatLR)
• generate - uses the same PCircuit architecture for P300 generation, independently for each subject
• both - demonstrates classification and generation

Check inside eeg_classify_generate.py for more details.

This required the addition of 2 solvers and fixed_point_quadratic primitive.

[toncho11]

I am going to work more about it first.