Feature extraction without learning using Hierarchical Temporal Memory

Abstract →

Hierarchical Temporal Memory (HTM) is a neuromorphic algorithm that emulates sparsity, hierarchy and modularity resembling the working principles of neocortex. Feature encoding is an important step to create sparse binary patterns. This sparsity is introduced by the binary weights and random weight assignment in the initialization stage of the HTM. We propose the alternative deterministic method for the HTM initialization stage, which connects the HTM weights to the input data and preserves natural sparsity of the input information. Further, we introduce the hardware implementation of the deterministic approach and compare it to the traditional HTM and existing hardware implementation. We test the proposed approach on the face recognition problem and show that it outperforms the conventional HTM approach.

It's not hard to outperform the "conventional HTM approach", because it does not work (yet). They tested face recognition performance on 3 very old datasets (AR, ORL, and Yale), and reported accuracy of 83%-86%. To provide some perspective, these datasets have been solved at least a decade ago (> 99% accuracy) [1]

But wait, have they built some cool analog hardware with memristors? Nope, just SPICE simulations...

So yeah, from Kazakhstan, the land of Borat, with love! (sorry, couldn't resist).

[1] http://www.coxlab.org/pdfs/ECCV_PintoDiCarloCox_2008.pdf