M2RU: Memristive Minion Recurrent Unit for On-Chip Continual Learning at the Edge

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Zyarah, Abdullah M., Kudithipudi, Dhireesha
Natura: Preprint
Pubblicazione: 2025
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866908732542156800
author Zyarah, Abdullah M.
Kudithipudi, Dhireesha
author_facet Zyarah, Abdullah M.
Kudithipudi, Dhireesha
contents Continual learning on edge platforms remains challenging because recurrent networks depend on energy-intensive training procedures and frequent data movement that are impractical for embedded deployments. This work introduces M2RU, a mixed-signal architecture that implements the minion recurrent unit for efficient temporal processing with on-chip continual learning. The architecture integrates weighted-bit streaming, which enables multi-bit digital inputs to be processed in crossbars without high-resolution conversion, and an experience replay mechanism that stabilizes learning under domain shifts. M2RU achieves 15 GOPS at 48.62 mW, corresponding to 312 GOPS per watt, and maintains accuracy within 5 percent of software baselines on sequential MNIST and CIFAR-10 tasks. Compared with a CMOS digital design, the accelerator provides 29X improvement in energy efficiency. Device-aware analysis shows an expected operational lifetime of 12.2 years under continual learning workloads. These results establish M2RU as a scalable and energy-efficient platform for real-time adaptation in edge-level temporal intelligence.
format Preprint
id arxiv_https___arxiv_org_abs_2512_17299
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle M2RU: Memristive Minion Recurrent Unit for On-Chip Continual Learning at the Edge
Zyarah, Abdullah M.
Kudithipudi, Dhireesha
Machine Learning
Artificial Intelligence
Emerging Technologies
Continual learning on edge platforms remains challenging because recurrent networks depend on energy-intensive training procedures and frequent data movement that are impractical for embedded deployments. This work introduces M2RU, a mixed-signal architecture that implements the minion recurrent unit for efficient temporal processing with on-chip continual learning. The architecture integrates weighted-bit streaming, which enables multi-bit digital inputs to be processed in crossbars without high-resolution conversion, and an experience replay mechanism that stabilizes learning under domain shifts. M2RU achieves 15 GOPS at 48.62 mW, corresponding to 312 GOPS per watt, and maintains accuracy within 5 percent of software baselines on sequential MNIST and CIFAR-10 tasks. Compared with a CMOS digital design, the accelerator provides 29X improvement in energy efficiency. Device-aware analysis shows an expected operational lifetime of 12.2 years under continual learning workloads. These results establish M2RU as a scalable and energy-efficient platform for real-time adaptation in edge-level temporal intelligence.
title M2RU: Memristive Minion Recurrent Unit for On-Chip Continual Learning at the Edge
topic Machine Learning
Artificial Intelligence
Emerging Technologies
url https://arxiv.org/abs/2512.17299