AI Poincare

STEP ONE: Select a dataset!

You could either (1) Choose from our library                            
                       or (2) Upload your favorite txt file:
{% if session.model %}

You are using the dataset:{{session.model}}.

{% else %}

You have not chosen any dataset.

{% endif %}

STEP TWO (Optional): Feel free to play with hyperparameters of AI Poincare!

A. Preprocessing: (1) Remove insignificant PCA components? {% if session.prepca == "yes" %} {% else %} {% endif %}     (2) Noise Threshold   

B. Fully-Connected Neural Network: (1) Hidden widths     (2) LeakyRelu slope     (3) L (length scale)    

       (4) Optimizer {% if session.opt == "Adam" %} {% else %} {% endif %}     (5) Learning Rate     (6) Batch Size     (7) Training Iterations

C. Measuring Neff: (1) a     (2) n_walk        

{% if save %}

You have successfully saved your preferences.

{% else %}

You are using default parameters.

{% endif %}

STEP THREE: Training the neural network and make a conclusion!

{% if update %} Number of conserved quantities: {{neff}} (Linear: {{remove_dim}}). Confidence Level: {{confidence}}.
For more details, please see the two figures we created for you: (1)Explained Ratio Diagram at "aipoincare/static/images/ERD.png" (2) Neff histogram at "aipoincare/static/images/Neff.png." {% else %} Neff is ???. Confidence Level is ???. {% endif %}