Strong Allee effect synaptic plasticity rule in an unsupervised learning environment (2203.13650v1)

Abstract: Synaptic plasticity or the ability of a brain to changes one or more of its functions or structures has generated and is sill generating a lot of interest from the scientific community especially neuroscientists. These interests especially went into high gear after empirical evidences were collected that challenged the established paradigm that human brain structures and functions are set from childhood and only modest changes were expected beyond. Early synaptic plasticity rules or laws to that regard include the basic Hebbian rule that proposed a mechanism for strengthening or weakening of synapses (weights) during learning and memory. This rule however did not account from the fact that weights must have bounded growth overtime. Thereafter, many other rules were proposed to complement the basic Hebbian rule and they also possess other desirable properties. In particular, a desirable property in synaptic plasticity rule is that the ambient system must account for inhibition which is often achieved if the rule used allows for a lower bound in synaptic weights. In this paper, we propose a synaptic plasticity rule inspired from the Allee effect, a phenomenon often observed in population dynamics. We show properties such such as synaptic normalization, competition between weights, de-correlation potential, and dynamic stability are satisfied. We show that in fact, an Allee effect in synaptic plasticity can be construed as an absence of plasticity.