The Concept of Metal-Insulator-Metal Nanostructures as Adaptive Neural Networks

Present computer processing capabilities are becoming a restriction to meet modern technological needs. Therefore, approaches beyond the von Neumann computational architecture are imperative and the brain operation and structure are truly attractive models. Memristors are characterized by a nonlinear relationship between current history and voltage and were shown to present properties resembling those of biological synapses. Here, the use of metal-insulator-metal-based memristive devices in neural networks capable of simulating the learning and adaptation features present in mammal brains is discussed.

Download Full-text

Improved Refractive Index-Sensing Performance of Multimode Fano-Resonance-Based Metal-Insulator-Metal Nanostructures

Nanomaterials ◽

10.3390/nano11082097 ◽

2021 ◽

Vol 11 (8) ◽

pp. 2097

Author(s):

Yuan-Fong Chou Chau ◽

Chung-Ting Chou Chao ◽

Siti Zubaidah Binti Haji Jumat ◽

Muhammad Raziq Rahimi Kooh ◽

Roshan Thotagamuge ◽

...

Keyword(s):

Refractive Index ◽

Plasmon Resonance ◽

Fano Resonance ◽

High Sensitivity ◽

Metal Nanostructures ◽

Refractive Index Sensor ◽

Metal Insulator ◽

Wide Range ◽

Metal Insulator Metal ◽

Mode 2

This work proposed a multiple mode Fano resonance-based refractive index sensor with high sensitivity that is a rarely investigated structure. The designed device consists of a metal–insulator–metal (MIM) waveguide with two rectangular stubs side-coupled with an elliptical resonator embedded with an air path in the resonator and several metal defects set in the bus waveguide. We systematically studied three types of sensor structures employing the finite element method. Results show that the surface plasmon mode’s splitting is affected by the geometry of the sensor. We found that the transmittance dips and peaks can dramatically change by adding the dual air stubs, and the light–matter interaction can effectively enhance by embedding an air path in the resonator and the metal defects in the bus waveguide. The double air stubs and an air path contribute to the cavity plasmon resonance, and the metal defects facilitate the gap plasmon resonance in the proposed plasmonic sensor, resulting in remarkable characteristics compared with those of plasmonic sensors. The high sensitivity of 2600 nm/RIU and 1200 nm/RIU can simultaneously achieve in mode 1 and mode 2 of the proposed type 3 structure, which considerably raises the sensitivity by 216.67% for mode 1 and 133.33% for mode 2 compared to its regular counterpart, i.e., type 2 structure. The designed sensing structure can detect the material’s refractive index in a wide range of gas, liquids, and biomaterials (e.g., hemoglobin concentration).

Download Full-text

Enhanced Harmonic Generation in Metal-Insulator-Metal Nanostructures

Conference on Lasers and Electro-Optics ◽

10.1364/cleo_at.2018.jth2a.71 ◽

2018 ◽

Author(s):

Mallik M. R. Hussain ◽

Zhengning Gao ◽

Domenico de Ceglia ◽

Maria A. Vincenti ◽

Andrew Sarangan ◽

...

Keyword(s):

Harmonic Generation ◽

Metal Nanostructures ◽

Metal Insulator ◽

Metal Insulator Metal

Download Full-text

Quantum conductivity for metal–insulator–metal nanostructures

Journal of the Optical Society of America B ◽

10.1364/josab.31.000259 ◽

2014 ◽

Vol 31 (2) ◽

pp. 259 ◽

Cited By ~ 44

Author(s):

Joseph W. Haus ◽

Domenico de Ceglia ◽

Maria Antonietta Vincenti ◽

Michael Scalora

Keyword(s):

Metal Nanostructures ◽

Metal Insulator ◽

Metal Insulator Metal

Download Full-text

Photoacoustic properties of metal-insulator-metal nanostructures

14th National Conference on Laser Technology and Optoelectronics (LTO 2019) ◽

10.1117/12.2534004 ◽

2019 ◽

Author(s):

Chunqi Zheng ◽

Huanzheng Zhu ◽

Jingyi Tian ◽

Qiang Li

Keyword(s):

Metal Nanostructures ◽

Metal Insulator ◽

Metal Insulator Metal

Download Full-text

Deep Neural Network Inverse-Design for Long Wave Infrared Hyperspectral Imaging

Applied Computational Electromagnetics Society ◽

10.47037/2020.aces.j.351137 ◽

2021 ◽

Vol 35 (11) ◽

pp. 1336-1337

Author(s):

Clayton Fowler ◽

Sensong An ◽

Bowen Zheng ◽

Hong Tang ◽

Hang Li ◽

...

Keyword(s):

Neural Networks ◽

Hyperspectral Imaging ◽

Electromagnetic Waves ◽

Deep Neural Networks ◽

Structural Parameters ◽

Inverse Design ◽

Metal Insulator ◽

Complex Interactions ◽

Metal Insulator Metal ◽

Long Wave Infrared

This paper presents a deep learning approach for the inverse-design of metal-insulator-metal metasurfaces for hyperspectral imaging applications. Deep neural networks are able to compensate for the complex interactions between electromagnetic waves and metastructures to efficiently produce design solutions that would be difficult to obtain using other methods. Since electromagnetic spectra are sequential in nature, recurrent neural networks are especially suited for relating such spectra to structural parameters.

Download Full-text

Over 1000-fold enhancement of upconversion luminescence using water-dispersible metal-insulator-metal nanostructures

Nature Communications ◽

10.1038/s41467-018-07284-w ◽

2018 ◽

Vol 9 (1) ◽

Cited By ~ 29

Author(s):

Ananda Das ◽

Chenchen Mao ◽

Suehyun Cho ◽

Kyoungsik Kim ◽

Wounjhang Park

Keyword(s):

Upconversion Luminescence ◽

Metal Nanostructures ◽

Metal Insulator ◽

Water Dispersible ◽

Metal Insulator Metal

Download Full-text

Hot plasmonic electron-driven catalytic reactions on patterned metal–insulator–metal nanostructures

Nanoscale ◽

10.1039/c7nr02805a ◽

2017 ◽

Vol 9 (32) ◽

pp. 11667-11677 ◽

Cited By ~ 4

Author(s):

Sun Mi Kim ◽

Changhwan Lee ◽

Kalyan C. Goddeti ◽

Jeong Young Park

Keyword(s):

Catalytic Activity ◽

Co Oxidation ◽

Catalytic Reactions ◽

Light Irradiation ◽

Metal Nanostructures ◽

Two Dimensional ◽

Metal Insulator ◽

Metal Insulator Metal ◽

2D Arrays

We fabricated two-dimensional (2D) arrays of metal–insulator–metal (MIM) plasmonic nanoislands designed to efficiently shuttle hot plasmonic electrons. These MIM nanostructures exhibit higher catalytic activity under light irradiation, revealing a significant impact on the catalytic activity for CO oxidation.

Download Full-text

Neuromodulated Dopamine Plastic Networks for Heterogeneous Transfer Learning with Hebbian Principle

Symmetry ◽

10.3390/sym13081344 ◽

2021 ◽

Vol 13 (8) ◽

pp. 1344

Author(s):

Arjun Magotra ◽

Juntae Kim

Keyword(s):

Neural Networks ◽

Transfer Learning ◽

Hebbian Learning ◽

Synaptic Connection ◽

Novel Approach ◽

Learning And Adaptation ◽

Algorithmic Technique ◽

Learning Principles ◽

The Brain ◽

Heterogeneous Source

The plastic modifications in synaptic connectivity is primarily from changes triggered by neuromodulated dopamine signals. These activities are controlled by neuromodulation, which is itself under the control of the brain. The subjective brain’s self-modifying abilities play an essential role in learning and adaptation. The artificial neural networks with neuromodulated plasticity are used to implement transfer learning in the image classification domain. In particular, this has application in image detection, image segmentation, and transfer of learning parameters with significant results. This paper proposes a novel approach to enhance transfer learning accuracy in a heterogeneous source and target, using the neuromodulation of the Hebbian learning principle, called NDHTL (Neuromodulated Dopamine Hebbian Transfer Learning). Neuromodulation of plasticity offers a powerful new technique with applications in training neural networks implementing asymmetric backpropagation using Hebbian principles in transfer learning motivated CNNs (Convolutional neural networks). Biologically motivated concomitant learning, where connected brain cells activate positively, enhances the synaptic connection strength between the network neurons. Using the NDHTL algorithm, the percentage of change of the plasticity between the neurons of the CNN layer is directly managed by the dopamine signal’s value. The discriminative nature of transfer learning fits well with the technique. The learned model’s connection weights must adapt to unseen target datasets with the least cost and effort in transfer learning. Using distinctive learning principles such as dopamine Hebbian learning in transfer learning for asymmetric gradient weights update is a novel approach. The paper emphasizes the NDHTL algorithmic technique as synaptic plasticity controlled by dopamine signals in transfer learning to classify images using source-target datasets. The standard transfer learning using gradient backpropagation is a symmetric framework. Experimental results using CIFAR-10 and CIFAR-100 datasets show that the proposed NDHTL algorithm can enhance transfer learning efficiency compared to existing methods.

Download Full-text