Online Arbitrary Shaped Clustering through Correlated Gaussian Functions

TL;DR

This paper introduces an online, unsupervised clustering algorithm that identifies arbitrarily shaped clusters without prior knowledge of cluster count, aiming for biological plausibility over traditional backpropagation.

Contribution

It presents a novel clustering method based on correlated Gaussian functions that can produce arbitrary shaped clusters and does not require pre-specified cluster numbers.

Findings

Works well on toy datasets across various hyperparameters

Produces arbitrary shaped clusters without prior cluster count

Offers a more biologically plausible alternative to backpropagation

Abstract

There is no convincing evidence that backpropagation is a biologically plausible mechanism, and further studies of alternative learning methods are needed. A novel online clustering algorithm is presented that can produce arbitrary shaped clusters from inputs in an unsupervised manner, and requires no prior knowledge of the number of clusters in the input data. This is achieved by finding correlated outputs from functions that capture commonly occurring input patterns. The algorithm can be deemed more biologically plausible than model optimization through backpropagation, although practical applicability may require additional research. However, the method yields satisfactory results on several toy datasets on a noteworthy range of hyperparameters.