An Online Algorithm for Learning Selectivity to Mixture Means

TL;DR

This paper introduces Triplet BCM, a biologically plausible online learning rule that converges to mixture model class means, extending classical BCM with triplet samples and linking it to tensor decomposition.

Contribution

The paper proposes Triplet BCM, a novel extension of BCM that uses triplet samples for improved mixture learning and provides theoretical convergence proofs.

Findings

Triplet BCM converges to mixture class means.

Provides a tensor decomposition interpretation of BCM.

Achieves a significant generalization over classical BCM.

Abstract

We develop a biologically-plausible learning rule called Triplet BCM that provably converges to the class means of general mixture models. This rule generalizes the classical BCM neural rule, and provides a novel interpretation of classical BCM as performing a kind of tensor decomposition. It achieves a substantial generalization over classical BCM by incorporating triplets of samples from the mixtures, which provides a novel information processing interpretation to spike-timing-dependent plasticity. We provide complete proofs of convergence of this learning rule, and an extended discussion of the connection between BCM and tensor learning.