Cognitive Active Medium: A Unified Framework for Ionic and Spin-Based Distributed Memory and a Metallic Neuromorphic Processor Architecture
Year: 2026
Type: article
Abstract: This preprint introduces the concept of a cognitive active medium (CAM) — a material that integrates memory, computation, and communication within a single physical substrate, eliminating the need for separate control layers. The framework unifies two material platforms: ion-polymer gels, where ionic migration enables distributed analog memory, and high-entropy alloy (HEA) spin glasses, where frustrated magnetic interactions support an exponentially large state space. A four-level architecture (mapping, addressing, cascading, learning) is proposed for a metallic neuromorphic processor. The work outlines a five‑phase experimental roadmap and identifies controlled cascade switching as the key scientific challenge. This preprint precedes a planned submission to Nature Electronics.
Author: Filonenko, Oleg
Language: English
Subfield: Electrical and Electronic Engineering
Field: Engineering
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W7225459634
