A Simple Reservoir Model of Working Memory with Real Values

TL;DR

This paper investigates how echo state networks can serve as simple models of working memory by storing and manipulating real values, highlighting the benefits of explicit memory in neural architectures.

Contribution

It introduces a simple reservoir model demonstrating how ESNs can store and utilize real-valued information for cognitive tasks, emphasizing the importance of explicit memory.

Findings

Explicit memory improves performance in storing real values.

Feedback connections enhance the network's computational capabilities.

Reservoir models can act as basic working memory units.

Abstract

The prefrontal cortex is known to be involved in many high-level cognitive functions, in particular, working memory. Here, we study to what extent a group of randomly connected units (namely an Echo State Network, ESN) can store and maintain (as output) an arbitrary real value from a streamed input, i.e. can act as a sustained working memory unit. Furthermore, we explore to what extent such an architecture can take advantage of the stored value in order to produce non-linear computations. Comparison between different architectures (with and without feedback, with and without a working memory unit) shows that an explicit memory improves the performances.