End of last year I spent a month as a guest researcher at Numenta, a machine intelligence company in Redwood City, CA – I mentioned them in an earlier post on sparse distributed representations. They recently started a visiting scholar…

In the present blog post I would like to present a slighlty altered version of the proof of a theorem in [1]. The statement is about the finite-sample expressivity of neural networks. You can basically take the approach and present it…

When reading about sequence learning and the prediction of elements in a time series you inevitably cross paths with the area of sequence compression at some point, and so did I. In the present post I’d like to outline a…