t-SNE Python implementation: Kullback-Leibler divergence

Question

t-SNE, as in [1], works by progressively reducing the Kullback-Leibler (KL) divergence, until a certain condition is met.
The creators of t-SNE suggests to use KL divergence as a performance criterion for the visualizations:

you can compare the Kullback-Leibler divergences that t-SNE reports. It is perfectly fine to run t-SNE ten times, and select the solution with the lowest KL divergence [2]

I tried two implementations of t-SNE:

• python: sklearn.manifold.TSNE().
• R: tsne, from library(tsne).

Both these implementations, when verbosity is set, print the error (Kullback-Leibler divergence) for each iteration. However, they don't allow the user to get this information, which looks a bit strange to me.

For example, the code:

import numpy as np
from sklearn.manifold import TSNE
X = np.array([[0, 0, 0], [0, 1, 1], [1, 0, 1], [1, 1, 1]])
model = TSNE(n_components=2, verbose=2, n_iter=200)
t = model.fit_transform(X)

produces:

[t-SNE] Computing pairwise distances...
[t-SNE] Computed conditional probabilities for sample 4 / 4
[t-SNE] Mean sigma: 1125899906842624.000000
[t-SNE] Iteration 10: error = 6.7213750, gradient norm = 0.0012028
[t-SNE] Iteration 20: error = 6.7192064, gradient norm = 0.0012062
[t-SNE] Iteration 30: error = 6.7178683, gradient norm = 0.0012114
...
[t-SNE] Error after 200 iterations: 0.270186

Now, as far as I understand, 0.270186 should be the KL divergence. However I cannot get this information, neither from model nor from t (which is a simple numpy.ndarray).

To solve this problem I could:

1. Calculate KL divergence by my self,
2. Do something nasty in python for capturing and parsing TSNE() function's output [3].

However:

1. would be quite stupid to re-calculate KL divergence, when TSNE() has already computed it,
2. would be a bit unusual in terms of code.

Do you have any other suggestion? Is there a standard way to get this information using this library?

I mentioned I tried R's tsne library, but I'd prefer the answers to focus on the python sklearn implementation.

---

References

[1] http://nbviewer.ipython.org/urls/gist.githubusercontent.com/AlexanderFabisch/1a0c648de22eff4a2a3e/raw/59d5bc5ed8f8bfd9ff1f7faa749d1b095aa97d5a/t-SNE.ipynb

[2] http://homepage.tudelft.nl/19j49/t-SNE.html

[3] https://stackoverflow.com/questions/16571150/how-to-capture-stdout-output-from-a-python-function-call

Answer 1

The TSNE source in scikit-learn is in pure Python. Fit fit_transform() method is actually calling a private _fit() function which then calls a private _tsne() function. That _tsne() function has a local variable error which is printed out at the end of the fit. Seems like you could pretty easily change one or two lines of source code to have that value returned to fit_transform().