Programmable memtransistor array with temporal dynamics modulation for efficient time-series data processing

Программируемый массив мемтранзисторов с модуляцией временной динамики для эффективной обработки данных временных рядов
Shinhyun Choi, Yujin Kim, Young Taek Oh, Dae‐won Kim, Yoonho Cho, Seokho Seo, See‐On Park, Taehwan Jang, Chaebin Park, Jae-Duk Lee
2026-07-04

Lorenz attractor forecastingmultiscale time-series processingprogrammable memtransistor arrayreservoir computingtemporal dynamics modulation
Hardware-based reservoir computing systems promise energy-efficient time-series data processing. However, their non-adjustable temporal dynamics at the hardware level limit multiscale feature extraction and computing capacity. Here, we present a programmable dynamic memtransistor featuring a dual-functional gate stack that enables hardware-level control of dynamic behavior without constant biasing or input preprocessing. The memtransistor integrates a charge storage layer for volatile temporal processing and a charge trap layer for non-volatile modulation of the band structure, enabling a 5-fold tunability in relaxation time constants. By configuring memtransistors with distinct dynamics in parallel, we realize a wide reservoir computing system capable of processing signals across multiple timescales. This architecture delivers a 40-fold reduction in the error for the multiple superimposed oscillator prediction task compared to a single reservoir baseline, and achieves software-comparable accuracy in forecasting the chaotic Lorenz attractor. Our results establish a compact and energy-efficient hardware platform for scalable wide reservoir computing implementation. Tunable temporal dynamics are needed for efficient hardware processing of time-series data. Here, the authors demonstrate a memtransistor array that non-volatilely tunes device dynamics for multiscale signal processing without extra circuitry.
1
A programmable dynamic memtransistor uses a dual-functional gate stack to control temporal dynamics at the hardware level without constant biasing or input preprocessing.
2
Parallel memtransistors with distinct dynamics form a wide reservoir computing system capable of processing signals across multiple timescales.
3
The hardware achieves software-comparable accuracy in forecasting the chaotic Lorenz attractor, supporting compact and energy-efficient scalable reservoir computing.
4
The system reduces prediction error 40-fold for multiple superimposed oscillator signals compared with a single-reservoir baseline.
5
Volatile charge storage enables temporal processing, while non-volatile charge trapping modulates the band structure and provides 5-fold tunability of relaxation time constants.

programmable dynamic memtransistor array with parallel memtransistors configured for reservoir computing

hardware-level non-volatile modulation and multiscale tunability of temporal dynamics for time-series processing, including relaxation-time control and forecasting performance

Publication Details
Publication Date
2026-07-04
Journal
Publisher
ISSN
Cited by
0
Access Type
Author Information
Authors
Shinhyun Choi
Yujin Kim
Young Taek Oh
Dae‐won Kim
Yoonho Cho
Seokho Seo
See‐On Park
Taehwan Jang
Chaebin Park
Jae-Duk Lee
Explore further
Open the scid.ai AI chat with a ready-made request: it will find papers on a similar topic and help build a literature review.
Find similar papers in the chat →
Make a presentation
100%