Concentric ESN: Assessing the Effect of Modularity in Cycle Reservoirs

TL;DR

This paper presents concentric Echo State Networks, a modular reservoir topology that improves predictive accuracy and memory capacity over traditional cycle and deep reservoir models.

Contribution

It introduces a novel concentric reservoir design that combines simple cycles with bidirectional jump connections, bridging simple and deep reservoir computing approaches.

Findings

Concentric reservoirs outperform single cycle reservoirs in accuracy.

Concentric reservoirs have higher memory capacity.

Preliminary results show improved predictive performance.

Abstract

The paper introduces concentric Echo State Network, an approach to design reservoir topologies that tries to bridge the gap between deterministically constructed simple cycle models and deep reservoir computing approaches. We show how to modularize the reservoir into simple unidirectional and concentric cycles with pairwise bidirectional jump connections between adjacent loops. We provide a preliminary experimental assessment showing how concentric reservoirs yield to superior predictive accuracy and memory capacity with respect to single cycle reservoirs and deep reservoir models.