Spatiotemporal Integration‐Capable Dual‐Gate Dendristor for Dendritic Computation and Image Processing
Authors:
Abstract
Dendritic computation based on the spatiotemporal integration for excitatory and inhibitory synaptic inputs plays a pivotal role in performing sophisticated information processing. It is vital to develop artificial dendritic devices with a flexible spatiotemporal integration capacity for excitatory and inhibitory inputs driven by the inputs with the same polarity. Here, a dual‐gate configuration dendristor (abbreviated as DGD) is demonstrated with p‐type and n‐type memtransistor characteristics based on Na+‐doped MoS2. The excitatory and inhibitory synaptic responses are achieved on the DGD under the same polarity voltage stimulations. Typical dendritic computational functions, including temporal coupling, cooperation and competition, and nonlinear integration, are successfully realized. The operational mechanisms are comprehensively elaborated. Benefitting from the flexible dendritic integration capacity, the DGD exhibits remarkable image denoise and dark/light adaptation capabilities. The recognition accuracy for a noisy image with 18% noise level is increased from 64% to 93% after preprocessing. Furthermore, the DGD has retina‐like dark/bright adaptation, increasing the recognition accuracies from 67% and 87% to 94% and 93% for low‐ and high‐light images, respectively. The DGD could provide a promising candidate for next‐generation edge intelligent devices capable of performing sophisticated spatiotemporal information processing tasks.Keywords
This publication has 68 references indexed in Scilit:
- BCM‐Inspired Synapses Constructed with Barrier‐Modulated Coupling Junctions for Enhancing Speech RecognitionAdvanced Functional Materials, 2024
- Direct Observation of Oxygen Ion Dynamics in a WO 3‐x based Second‐Order Memristor with Dendritic Integration FunctionsAdvanced Functional Materials, 2023
- Multifunctional Optoelectronic Synapses Based on Arrayed MoS2 Monolayers Emulating Human Association MemoryAdvanced Science, 2023
- A Memristors‐Based Dendritic Neuron for High‐Efficiency Spatial‐Temporal Information ProcessingAdvanced Materials, 2022
- Two‐Inch Wafer‐Scale Exfoliation of Hexagonal Boron Nitride Films Fabricated by RF‐SputteringAdvanced Functional Materials, 2022
- Sodium‐Doped Titania Self‐Rectifying Memristors for Crossbar Array Neuromorphic ArchitecturesAdvanced Materials, 2021
- Grain‐Boundary Engineering of Monolayer MoS2 for Energy‐Efficient Lateral Synaptic DevicesAdvanced Materials, 2021
- Controlling Nanoscale Thermal Expansion of Monolayer Transition Metal Dichalcogenides by Alloy EngineeringSmall, 2019
- From Bulk to Monolayer MoS2: Evolution of Raman ScatteringAdvanced Functional Materials, 2012
- Low‐Work‐Function Surface Formed by Solution‐Processed and Thermally Deposited Nanoscale Layers of Cesium CarbonateAdvanced Functional Materials, 2007