Neuromorphic computing based on halide perovskites

Nature Electronics, Journal Year: 2023, Volume and Issue: 6(12), P. 949 - 962

Published: Dec. 21, 2023

Language: Английский

---

Highly-scaled and fully-integrated 3-dimensional ferroelectric transistor array for hardware implementation of neural networks

Nature Communications, Journal Year: 2023, Volume and Issue: 14(1)

Published: Jan. 31, 2023

Abstract Hardware-based neural networks (NNs) can provide a significant breakthrough in artificial intelligence applications due to their ability extract features from unstructured data and learn them. However, realizing complex NN models remains challenging because different tasks, such as feature extraction classification, should be performed at memory elements arrays. This further increases the required number of arrays chip size. Here, we propose three-dimensional ferroelectric NAND (3D FeNAND) array for area-efficient hardware implementation NNs. Vector-matrix multiplication is successfully demonstrated using integrated 3D FeNAND arrays, excellent pattern classification achieved. By allocating each vertical layers hidden layer NN, used perform color-mixed patterns work provides practical strategy realize high-performance highly efficient systems by stacking computation components vertically.

Language: Английский

---

Functional Materials for Memristor‐Based Reservoir Computing: Dynamics and Applications

Advanced Functional Materials, Journal Year: 2023, Volume and Issue: 33(42)

Published: June 23, 2023

Abstract The booming development of artificial intelligence (AI) requires faster physical processing units as well more efficient algorithms. Recently, reservoir computing (RC) has emerged an alternative brain‐inspired framework for fast learning with low training cost, since only the weights associated output layers should be trained. Physical RC becomes one leading paradigms computation using high‐dimensional, nonlinear, dynamic substrates. Among them, memristor appears to a simple, adaptable, and constructing they exhibit nonlinear features memory behavior, while memristor‐implemented neural networks display increasing popularity towards neuromorphic computing. In this review, systems from following aspects: architectures, materials, applications are summarized. It starts introduction structures that can simulated blocks. Specific interest then focuses on behaviors memristors based various material systems, optimizing understanding relationship between relaxation which provides guidance references building coped on‐demand application scenarios. Furthermore, recent advances in memristor‐based surveyed. end, further prospects system view envisaged.

Language: Английский

---

Reconfigurable Neuromorphic Computing: Materials, Devices, and Integration

Advanced Materials, Journal Year: 2023, Volume and Issue: 35(51)

Published: June 7, 2023

Abstract Neuromorphic computing has been attracting ever‐increasing attention due to superior energy efficiency, with great promise promote the next wave of artificial general intelligence in post‐Moore era. Current approaches are, however, broadly designed for stationary and unitary assignments, thus encountering reluctant interconnections, power consumption, data‐intensive that domain. Reconfigurable neuromorphic computing, an on‐demand paradigm inspired by inherent programmability brain, can maximally reallocate finite resources perform proliferation reproducibly brain‐inspired functions, highlighting a disruptive framework bridging gap between different primitives. Although relevant research flourished diverse materials devices novel mechanisms architectures, precise overview remains blank urgently desirable. Herein, recent strides along this pursuit are systematically reviewed from material, device, integration perspectives. At material device level, one comprehensively conclude dominant reconfigurability, categorized into ion migration, carrier phase transition, spintronics, photonics. Integration‐level developments reconfigurable also exhibited. Finally, perspective on future challenges is discussed, definitely expanding its horizon scientific communities.

Language: Английский

---

Programming memristor arrays with arbitrarily high precision for analog computing

Science, Journal Year: 2024, Volume and Issue: 383(6685), P. 903 - 910

Published: Feb. 22, 2024

In-memory computing represents an effective method for modeling complex physical systems that are typically challenging conventional architectures but has been hindered by issues such as reading noise and writing variability restrict scalability, accuracy, precision in high-performance computations. We propose demonstrate a circuit architecture programming protocol converts the analog result to digital at last step enables low-precision devices perform high-precision computing. use weighted sum of multiple represent one number, which subsequently programmed used compensate preceding errors. With memristor system-on-chip, we experimentally solutions scientific tasks while maintaining substantial power efficiency advantage over approaches.

Language: Английский

---

Solution-processed memristors: performance and reliability

Nature Reviews Materials, Journal Year: 2024, Volume and Issue: 9(5), P. 358 - 373

Published: April 12, 2024

Language: Английский

---

Single-Pore Nanofluidic Logic Memristor with Reconfigurable Synaptic Functions and Designable Combinations

Journal of the American Chemical Society, Journal Year: 2024, Volume and Issue: 146(21), P. 14558 - 14565

Published: May 16, 2024

The biological neural network is a highly efficient in-memory computing system that integrates memory and logical functions within synapses. Moreover, reconfiguration by environmental chemical signals endows networks with dynamic multifunctions enhanced efficiency. Nanofluidic memristors have emerged as promising candidates for mimicking synaptic functions, owing to their similarity synapses in the underlying mechanisms of ion signaling channels. However, realizing signal-modulated logic nanofluidic memristors, which basis brain-like applications, remains unachieved. Here, we report single-pore memristor reconfigurable functions. Based on different degrees protonation deprotonation functional groups inner surface single pore, modulation are realized. More noteworthy, this can not only avoid average effects multipore but also act fundamental component constructing complex through series parallel circuits, lays groundwork future artificial networks. implementation gates signals, diverse combinations opens up new possibilities applications brain-inspired computing.

Language: Английский

---

Inductive and Capacitive Hysteresis of Current-Voltage Curves: Unified Structural Dynamics in Solar Energy Devices, Memristors, Ionic Transistors, and Bioelectronics

PRX Energy, Journal Year: 2024, Volume and Issue: 3(1)

Published: Jan. 8, 2024

Hysteresis observed in the current-voltage curves of both electronic and ionic devices is a phenomenon where curve's shape altered on basis measurement speed. This effect driven by internal processes that introduce time delay response to an external stimulus, leading measurements being dependent history past disturbances. hysteresis has posed challenges, particularly solution-processed photovoltaic such as halide perovskite solar cells, it significantly complicates evaluation performance quality. In other devices, memristors organic electrochemical transistors for neuromorphic applications, inherent aspect their functionality, facilitating transitions between different conductivity states. Natural artificial ionically conducting channels also exhibit pronounced hysteresis, crucial component generating action potentials neurons. this study, we aim categorize various forms identifying shared elements among diverse physical, chemical, biological systems. Our method involves examining from multiple angles, using simplified models capture essential types. We analyze system behavior techniques linear sweep voltammetry impedance spectroscopy transient currents resulting small voltage steps. investigation reveals two primary types based how current responds rapid rates: capacitive inductive hysteresis. These terms correspond dominant equivalent circuit, determining response. Remarkably, these concepts provide insights into vastly systems, spanning capacitors, transistors, electrofluidic nanopores, protein ion channels. The consistency electrical responses across cases enables identification cause elucidate frequency dependence stepwise illustrating fundamental relaxations contribute overall surplus or deficit during extensive sweeps define curve. Published American Physical Society 2024

Language: Английский

---

Biomaterial/Organic Heterojunction Based Memristor for Logic Gate Circuit Design, Data Encryption, and Image Reconstruction

Advanced Functional Materials, Journal Year: 2024, Volume and Issue: 34(36)

Published: March 13, 2024

Abstract Benefiting from powerful logic‐computing, higher packaging density, and extremely low electricity consumption, memristors are regarded as the most promising next‐generation of electric devices capable realizing brain‐like neuromorphic computation. However, design emerging circuit based on their potential application in unconventional fields very meaningful for achieving some tasks that traditional electronic cannot accomplish. Herein, a Cu/PEDOT:PSS‐PP:PVDF/Ti structured memristor is fabricated by using polyvinylidene difluoride (PVDF) dopped biomaterial papaya peel (PP) organic poly(3,4‐ethylenedioxythiophene):polystyrene sulfonate (PEDOT:PSS) heterojunction functional layer, which can be switched among resistive switching, self‐rectification effect, capacitive behavior adjusting voltage bias/scan rate. Through further fitting data simulating interfacial group reactions, this work innovatively proposes charge conduction mode device driven Fowler–Nordheim tunneling, complexation PEDOT:PSS pore removal. Finally, regular logic gate adder circuits constructed memristor, while fully adder‐based encryption unit designed to realize image reconstruction. This renders compatible with circuits, widening path toward information security.

Language: Английский

---

Lead‐Free Perovskites and Metal Halides for Resistive Switching Memory and Artificial Synapse

Small Structures, Journal Year: 2024, Volume and Issue: 5(6)

Published: April 3, 2024

Memristive devices such as resistive switching memories and artificial synapses have emerged promising technologies to overcome the technological challenges associated with von Neumann bottleneck. Recently, lead halide perovskites drawn substantial research attention candidate material for memristors due their unique optoelectronic properties, solution processability, mechanical flexibility. However, toxicity of lead‐containing species has raised major concerns health environment, which makes it crucial transition from lead‐based lead‐free materials practical applications. Herein, recent progress metal halides including perovskite analogs memory synapse is reviewed. Initially, fundamentals mechanisms are introduced. Next, design, fabrication technique, device performance summarized critically evaluated each species. Finally, toward outlined discussed, some potential directions future study proposed.

Language: Английский

---