Fully Analog Resonant Recurrent Neural Network via Metacircuit

arXiv cs.LG / 4/21/2026

📰 NewsDeveloper Stack & InfrastructureSignals & Early TrendsModels & Research

共有:

Key Points

The paper proposes a fully analog resonant recurrent neural network (R²NN) for temporal information processing to enable faster, more energy-efficient edge inference than conventional digital approaches.
It uses a metacircuit architecture of coupled electrical local resonators, along with a reformulated mechanical-electrical analogy, to directly map a trained R²NN model onto physical hardware elements accurately.
The system introduces jointly trainable global resistive coupling and local resonances that create effective frequency-dependent negative resistances, shaping an impedance landscape that routes currents through frequency-selective paths.
Because discriminative spectral features can be extracted directly from raw analog inputs, the approach supports real-time temporal classification while avoiding analog-to-digital conversion.
The authors demonstrate cross-domain hardware implementations for tactile perception, speech recognition, and condition monitoring, positioning the method as a scalable analog paradigm for intelligent edge processing.

Abstract

Physical neural networks offer a transformative route to edge intelligence, providing superior inference speed and energy efficiency compared to conventional digital architectures. However, realizing scalable, end-to-end, fully analog recurrent neural networks for temporal information processing remains challenging due to the difficulty of faithfully mapping trained network models onto physical hardware. Here we present a fully analog resonant recurrent neural network (R

^2

NN) implemented via a metacircuit architecture composed of coupled electrical local resonators. A reformulated mechanical-electrical analogy establishes a direct mapping between the R

^2

NN model and metacircuit elements, enabling accurate physical implementation of trained neural network parameters. By integrating jointly trainable global resistive coupling and local resonances, which generate effective frequency-dependent negative resistances, the architecture shapes an impedance landscape that steers currents along frequency-selective pathways. This mechanism enables direct extraction of discriminative spectral features, facilitating real-time temporal classification of raw analog inputs while bypassing analog-to-digital conversion. We demonstrate the cross-domain versatility of this framework using integrated hardware for tactile perception, speech recognition, and condition monitoring. This work establishes a scalable, fully analog paradigm for intelligent temporal processing and paves the way for low-latency, resource-efficient physical neural hardware for edge intelligence.