Neuromorphic one-shot learning utilizing a phase-transition material DOI
Alessandro R. Galloni, Yifan Yuan, Minning Zhu

et al.

Proceedings of the National Academy of Sciences, Journal Year: 2024, Volume and Issue: 121(17)

Published: April 17, 2024

Design of hardware based on biological principles neuronal computation and plasticity in the brain is a leading approach to realizing energy- sample-efficient AI learning machines. An important factor selection building blocks identification candidate materials with physical properties suitable emulate large dynamic ranges varied timescales signaling. Previous work has shown that all-or-none spiking behavior neurons can be mimicked by threshold switches utilizing material phase transitions. Here, we demonstrate devices prototypical metal-insulator-transition material, vanadium dioxide (VO 2 ), dynamically controlled access continuum intermediate resistance states. Furthermore, timescale their intrinsic relaxation configured match range biologically relevant from milliseconds seconds. We exploit these device three aspects analog computation: fast (~1 ms) soma compartment, slow (~100 dendritic ultraslow s) biochemical signaling involved temporal credit assignment for recently discovered mechanism one-shot learning. Simulations show an artificial neural network using VO control agent navigating spatial environment learn efficient path reward up fourfold fewer trials than standard methods. The relaxations described our study may engineered variety thermal, electrical, or optical stimuli, suggesting further opportunities neuromorphic hardware.

Language: Английский

Leaky integrate-and-fire and oscillation neurons based on ZnO diffusive memristors for spiking neural networks DOI
Liang Wang, Le Zhang,

Shuai‐Bin Hua

et al.

Science China Materials, Journal Year: 2025, Volume and Issue: unknown

Published: Jan. 24, 2025

Language: Английский

Citations

1
Toward Cognitive Machines: Evaluating Single Device Based Spiking Neural Networks for Brain-Inspired Computing DOI
Faisal Bashir, Ali Alzahrani, Haider Abbas

et al.

ACS Applied Electronic Materials, Journal Year: 2025, Volume and Issue: unknown

Published: Feb. 14, 2025

A brain-inspired computing paradigm known as "neuromorphic computing" seeks to replicate the information processing processes of biological neural systems in order create that are effective, low-power, and adaptable. Spiking networks (SNNs) based on a single device at forefront computing, which aims mimic powers human brain. Neuromorphic devices, enable hardware implementation artificial (ANNs), heart neuromorphic computing. These devices dynamics functions neurons synapses. This mini-review assesses latest advancements with an emphasis small, energy-efficient synapses neurons. Key like spike-timing-dependent plasticity, multistate storage, dynamic filtering demonstrated by variety single-device models, such memristors, transistors, magnetic ferroelectric devices. The integrate-and-fire (IF) neuron is key model these because it allows for mathematical analysis while successfully capturing aspects processing. review examines potential SNNs scalable, low-power applications, highlighting both benefits constraints implementing them architectures. highlights increasing importance creation flexible cognitive

Language: Английский

Citations

1
Bifurcation and oscillations in fluidic nanopores: A model neuron for liquid neuromorphic networks DOI Creative Commons
Alicia Cordero, Juan R. Torregrosa, Juan Bisquert

et al.

Physical Review Research, Journal Year: 2025, Volume and Issue: 7(1)

Published: March 18, 2025

Neuron spiking constitutes the central information node in neural networks. Nanoscale fluidic pores with rectification and hysteresis provide opportunity to induce voltage oscillations same physical principles as living neurons. We establish conditions that enable self-sustained limit-cycle a single artificial pore channel by Hopf bifurcation, thus providing minimal model for elementary neuron. On nanochannel contains necessary ingredients capacitive inductive response stationary negative resistance, we identify range of parameters where occur. These results crucial guidelines identifying build oscillating nanopore according system's geometrical, electrical, fluidic, chemical variables, which otherwise could be overlooked, since relevant parameter region producing can narrow. Published American Physical Society 2025

Language: Английский

Citations

1
Hardware Implementation of Network Connectivity Relationships Using 2D hBN‐Based Artificial Neuron and Synaptic Devices DOI Creative Commons
Yooyeon Jo, Dong Yeon Woo, Gichang Noh

et al.

Advanced Functional Materials, Journal Year: 2023, Volume and Issue: 34(10)

Published: Nov. 5, 2023

Abstract Brain‐inspired neuromorphic computing has been developed as a potential candidate for solving the von Neumann bottleneck of traditional systems. 2D materials‐based memristors have exponentially investigated promising building blocks because their excellent electrical performance, simple structure, and small device scale. However, while many researchers focused on looking into individual artificial devices based memristors, only few studies integration neuron synaptic reported. In this work, both volatile nonvolatile are fabricated by using hexagonal boron nitride film devices, respectively. The leaky‐integrate‐and‐fire performance functions (e.g., weight plasticity spike‐timing‐dependent plasticity) well emulated with devices. MNIST image classification is conducted experimental data. For first time, an neuron‐synapse‐neuron neural network physically constructed to mimic biological networks. connection strength modulation experimentally demonstrated between neurons depending conductance state synapse, paving way development large‐scale hardware.

Language: Английский

Citations

20
Multilevel Conductance States of Vapor‐Transport‐Deposited Sb2S3 Memristors Achieved via Electrical and Optical Modulation DOI Creative Commons
Somnath S. Kundale, Pravin S. Pawar, Dhananjay D. Kumbhar

et al.

Advanced Science, Journal Year: 2024, Volume and Issue: 11(32)

Published: July 3, 2024

The pursuit of advanced brain-inspired electronic devices and memory technologies has led to explore novel materials by processing multimodal multilevel tailored conductive properties as the next generation semiconductor platforms, due von Neumann architecture limits. Among such materials, antimony sulfide (Sb

Language: Английский

Citations

7
Reconfigurable Neuromorphic Computing with 2D Material Heterostructures for Versatile Neural Information Processing DOI
Jiayang Hu, Hanxi Li, Yishu Zhang

et al.

Nano Letters, Journal Year: 2024, Volume and Issue: 24(30), P. 9391 - 9398

Published: July 22, 2024

Reconfigurable neuromorphic computing holds promise for advancing energy-efficient neural network implementation and functional versatility. Previous work has focused on emulating specific functions rather than an integrated approach. We propose all two-dimensional (2D) material-based heterostructure capable of performing multiple operations by reconfiguring output terminals in response to stimuli. Specifically, our device can synergistically emulate the key elements synapse, neuron, dendrite, which play important interrelated roles information processing. Dendrites, branches that receive transmit presynaptic action potentials, possess ability nonlinearly integrate filter incoming signals. The proposed allows reconfiguration between different operation modes, demonstrating its potential diverse tasks. As a proof concept, we show perform basic Boolean logic functions. This highlights applicability complex neural-network-based processing problems. Our approach may advance development versatile, low-power hardware.

Language: Английский

Citations

6
Device physics recipe to make spiking neurons DOI Open Access
Juan Bisquert

Chemical Physics Reviews, Journal Year: 2023, Volume and Issue: 4(3)

Published: Sept. 1, 2023

Neurons, which are made of biological tissue, exhibit cognitive properties that can be replicated in various material substrates. To create brain-inspired computational artificial systems, we construct microscopic electronic neurons mimic natural systems. In this paper, discuss the essential and device needed for a spiking neuron, characterized using impedance spectroscopy small perturbation equivalent circuit elements. We find minimal neuron system requires capacitor, chemical inductor, negative resistance. These components integrated naturally physical response device, instead built from separate identify structural conditions smooth oscillations depend on certain dynamics conducting with internal state variables. variables diverse nature, such as fluids, solids, or ionic organic materials, implying functional ways. highlight importance detecting Hopf bifurcation, critical point achieving behavior, through spectral features impedance. end, provide systematic method analysis terms characteristic frequencies obtained methods. Thus, propose methodology to quantify devices produce dynamic necessary specific sensory-cognitive tasks. By replicating it may possible systems enhanced capabilities information processing, pattern recognition, learning. Additionally, understanding contribute our knowledge how function interact complex neural networks. Overall, paper presents novel approach toward building provides insight into important considerations behavior neurons.

Language: Английский

Citations

15
Artificial sensory system based on memristive devices DOI Creative Commons

Ju Young Kwon,

Ji Eun Kim,

Jong Sung Kim

et al.

Exploration, Journal Year: 2023, Volume and Issue: 4(1)

Published: Nov. 20, 2023

Abstract In the biological nervous system, integration and cooperation of parallel system receptors, neurons, synapses allow efficient detection processing intricate disordered external information. Such systems acquire process environmental data in real‐time, efficiently handling complex tasks with minimal energy consumption. Memristors can mimic typical by implementing key features neuronal signal‐processing functions such as selective adaption leaky integrate‐and‐fire synaptic plasticity synapses. External stimuli are sensitively detected filtered “artificial receptors,” encoded into spike signals via neurons,” integrated stored through synapses.” The high operational speed, low power consumption, superior scalability memristive devices make their high‐performance sensors a promising approach for creating artificial sensory systems. These extract useful from large volume raw data, facilitating real‐time This review explores recent advances memristor‐based authors begin requirements elements then present an in‐depth demonstrated devices. Finally, major challenges opportunities development discussed.

Language: Английский

Citations

14
Homochiral Chemistry Strategy to Trigger Second-Harmonic Generation and Dual Dielectric Switches DOI
Bowen Deng,

Zhu Yang,

Kun Ding

et al.

Inorganic Chemistry, Journal Year: 2023, Volume and Issue: 62(29), P. 11701 - 11707

Published: July 10, 2023

Switchable materials have attracted enormous interest due to their promising applications in important fields such as sensing, electronic components, and information storage. Nevertheless, obtaining multifunctional switching is still a problem worth investigating. Herein, by incorporating (Rac-, L-, D-2-amino-1-propanol) the templating cation, we obtained D-HTMPA)CdCl3 (HTMPA = 1-hydroxy-N, N, N-trimethyl-2-propanaminium). We adopted chiral chemistry strategy that causes (Rac-HTMPA)CdCl3 central symmetric space crystallize group. Based on modulation of homochiral strategy, (L-, shows dual phasic transition at 269 326 K switchable second-harmonic generation response. In addition, material exhibit stable dielectric (SHG) switches. This work provides an approach exploring materials.

Language: Английский

Citations

13
Filament-free memristors for computing DOI Creative Commons
Sanghyeon Choi, Taehwan Moon, Gunuk Wang

et al.

Nano Convergence, Journal Year: 2023, Volume and Issue: 10(1)

Published: Dec. 19, 2023

Memristors have attracted increasing attention due to their tremendous potential accelerate data-centric computing systems. The dynamic reconfiguration of memristive devices in response external electrical stimuli can provide highly desirable novel functionalities for applications when compared with conventional complementary-metal-oxide-semiconductor (CMOS)-based devices. Those most intensively studied and extensively reviewed memristors the literature so far been filamentary type memristors, which typically exhibit a relatively large variability from device switching cycle cycle. On other hand, filament-free shown better uniformity attractive dynamical properties, enable variety new paradigms but rarely reviewed. In this article, wide range corresponding are Various junction structures, principles surveyed discussed. Furthermore, we introduce recent advances different schemes demonstrations based on non-filamentary memristors. This Review aims present valuable insights guidelines regarding key computational primitives implementations enabled by these

Language: Английский

Citations

13