Metal Oxide Neural Devices and Their Applications

Highlights in Science Engineering and Technology, Journal Year: 2024, Volume and Issue: 87, P. 226 - 231

Published: March 26, 2024

Metallic oxide neurons are a potential candidate for future applications in neuroscience and neuroscience, where they can be used as model simulating synaptic functions biology. This paper mainly introduces two kinds of oxide-based memristors: memristor neural transistor, well briefly discuss their integration. Because the devices’ simple structure, similar to high efficiency, low power consumption, high-density integration, compatibility with CMOS process conditions, continuous control conductance, it is considered first choice brain-like computing hardware. And realize integration storage single device. However, some problems challenges still exist application. In device, there problem insufficient miniaturization, because influence leakage current, which limits array increases operation energy consumption. conclusion, despite significant barriers before becoming commercially viable, promising certain.

Language: Английский

---

Proposal and detailed theoretical analysis on photonic neural network with optically pumped Spin-VCSEL spiking neurons

Japanese Journal of Applied Physics, Journal Year: 2024, Volume and Issue: 63(9), P. 092004 - 092004

Published: Sept. 2, 2024

Abstract A dual-layer photonic spiking neural network (PSNN) was constructed, where multiple optically pumped spin vertical-cavity surface-emitting lasers (Spin-VCSELs) were proposed as neurons. Based on a detailed theoretical analysis of leaky integrate-and-fire (LIF) and refractory period characteristics Spin-VCSEL neurons, the training testing performance for studied PSNN evaluated using two standard pattern classification tasks (Iris dataset, simple digit recognition). The results showed that, by selecting appropriate parameters such frequency detuning number pre-synaptic etc., higher training/testing accuracies beyond 90% can be obtained. When compared with traditional electrically VCSEL, threshold reduction up to 50% achieved under nanosecond scale relaxation time circular polarization optical pumping, feasibility realizing high accuracy (88%) near reduced also verified. Therefore, neurons become valuable new choice high-performance power consumption.

Language: Английский

---

Programmable Optical Synaptic Linking of Neuromorphic Photonic-Electronic RTD Spiking Circuits

ACS Photonics, Journal Year: 2024, Volume and Issue: unknown

Published: Sept. 30, 2024

Interconnectivity between functional building blocks (such as neurons and synapses) represents a fundamental functionality for realizing neuromorphic systems. However, in the domain of photonics, synaptic interlinking cascadability spiking optical artificial remains challenging mostly unexplored experiments. In this work, we report an link optoelectronic based upon resonant tunneling diodes (RTDs) that allows cascadable spike propagation. First, deterministic is triggered using multimodal (electrical optical) inputs RTD-based neurons, which are (OE) circuits incorporating either micron-scale RTDs or photosensitive nanopillar-based RTDs. Second, feedforward linking with dynamical weighting signals pre- postsynaptic RTD demonstrated, including cascaded activation. By dynamically amplitude spikes, it shown how activation probability node directly follows weighted spikes. This work therefore provides first experimental demonstration programmable cascading multiple fast efficient OE providing key photonic-electronic neural networks light-enabled hardware.

Language: Английский

---

NeuroMIMO: Employing the Neuromorphic Computing Principles to Achieve Power-Efficient MU-MIMO Detection

Published: Nov. 11, 2024

Language: Английский

---