Intelligent Nanotechnology: Merging Nanoscience and Artificial Intelligence

Cover
Front Matter
Copyright
Contributors
Preface
Acknowledgments
Inverse design meets nanophotonics: From computational optimization to artificial neural network
Computational inverse design
Gradient-based
Adjoint optimization
Topology optimization
Evolutionary algorithms
Genetic algorithm
Particle swarm optimization
Deep learning-based inverse design
Deterministic neural network-based inverse design
Multilayer perceptron
Convolutional neural networks
Generative neural network-based inverse design
Autoencoders
Generative adversarial networks
Deep learning merged with computational optimization
Generative neural networks combined with topology optimization
Generative neural networks combined with metaheuristic optimization
References
Machine learning for solid mechanics
Introduction
Case studies
Surrogate modeling for materials
Case study: Prediction of material properties of binary composites
Examples of surrogate models
Inverse material design
Physics-informed neural network
Graph neural networks
Future opportunities and considerations
Conclusions
References
Neural networks in phononics
Introduction
One-dimensional phononic crystals and transfer matrix method
Forward prediction of phononic crystals with neural networks
Predicting the dispersion curves of one-dimensional phononic crystals
Datasets for the prediction of dispersion curves
Neural networks predicting dispersion curves
Choosing suitable neural networks
Predicted dispersion curves in the testing sets
Predicting the energy transmission spectrums of one-dimensional phononic crystals
Datasets for the prediction of energy transmission spectrums
Neural networks predicting energy transmission spectrums
Predicted energy transmission spectrums in the testing sets
Designing phononic crystals with neural networks
Datasets for the inverse design of phononic crystals
One-parameter design
Two-parameter design
Three-parameter design
Summary
References
Nanophotonic devices based on optimization algorithms
Introduction
Gradient-based algorithms
Topology optimization
Variable density method
Introduction
Applications
Level set method
Introduction
Applications
Bi-directional evolutionary structural optimization
Introduction
Applications
Objective-first algorithm
Introduction
Applications
Other gradient-based algorithms
Introduction
Applications
Heuristic algorithms
Hill-climbing algorithm
Introduction
Applications
Direct binary search
Introduction
Applications
Simulated annealing algorithm
Introduction
Applications
Tabu search
Introduction
Applications
Genetic algorithm
Introduction
Applications
Differential evolution algorithm
Introduction
Applications
Particle swarm optimization algorithm
Introduction
Applications
Ant colony algorithm
Introduction
Applications
Conclusion
References
Artificial intelligence (AI) enhanced nanomotors and active matter
Introduction
Synthetic active particles
Control of active particles
Light-controllable active particles
Feedback control of active particles
Experimental realization of feedback control
Reinforcement learning
Reinforcement learning with a real microswimmer
Deep reinforcement learning with a single active particle
Multi-agent systems
Multi-agent reinforcement learning (MARL)
Future directions for MARL
References
Applications of convolutional neural networks for spectral analysis
Introduction
Fundamentals of CNNs for photonics
Parameterization strategies
Mathematical operations
Sampling and training
Predictive models for spectra calculation
Fundamental spectra analysis
Ultrafast spectroscopy signal analysis
Generative models for spectra design
Generative CNNs
Applications of GCNN
Dimensionality reduction models for optical property extraction
Unsupervised learning with dimensionality reduction models
Applications of dimensionality reduction models
Perspectives and outlooks
References
Nanoscale electronic synapses for neuromorphic computing
Introduction
Realization of artificial synapses
Ion migration
Cation migration
Anion migration
Electrolyte-gated transistor
Electronic migration
Electron migration by external field
Photogenerated carrier migration
Phase transition
Ferroelectric
Magnetic
Realization of neuromorphic engineering
Synaptic plasticity
Long-term potentiation and depression (LTP/LTD)
Short-term potentiation and depression (STP/STD)
Spiking-time dependent plasticity (STDP)
Other plasticity
Integration and cognitive functions
Computing with synapses
Artificial neuromorphic devices
Summary
References
Nanowire memristor as artificial synapse in random networks
Introduction
Fundaments of NW-based memristive devices
NW synthesis
Fabrication strategies of NW-based memristive devices
ECM mechanism
VCM mechanism
Single nanowire memristor as artificial synapse
Nanowire random networks as artificial neural networks
Nanowire network topology
Memristive behavior of single network elements
``Reweighting´´ effect in single NW junctions
``Rewiring´´ effects in single NWs
Emergent dynamics
Synaptic plasticity in NW networks
Short-term and long-term synaptic plasticity
Structural plasticity
Homosynaptic and heterosynaptic plasticity
Criticality and avalanches effects
Computing with nanowire random networks
Conclusions
References
Artificial intelligence accelerator using photonic computing
Introduction
Optical weighted interconnections
Free-space optical weighted interconnection
Integrated optical weighted interconnection
Optical neuron activation functions
Photorefractive effect
Saturable absorption
Optical Kerr effect
Structural phase transition
Optical resonance
Photoelectronic effect
Designing photonic neural network architectures
Photonic fully connected neural networks
Photonic convolutional neural networks
Photonic recurrent neural networks
Photonic spiking neural networks
Optoelectronic devices and AI systems
On-chip light source
Photonic memory
Optoelectronic modulator
Optoelectronic AI systems
In-situ optical backpropagation training methods
Gradient calculation based on optical backpropagation
Gradient calculation based on electronic backpropagation
Discussion and outlook
References
Machine learning in nanomaterial electron microscopy data analysis
Introduction
High-throughput analysis of nanomaterial microscopy data
Machine learning models relevant to microscopy data analysis
Training datasets in supervised learning of nanomaterial microscopy images
ML in 2D microscopy image analysis
Classification and regression of nanomaterial microscopy images
Segmentation of nanomaterial microscopy images
Object detection of nanomaterial microscopy images
ML in 3D tomography reconstruction and segmentation
ML-assisted analysis of nonimage data
Conclusion and outlook
References
Deep learning in biomedical informatics
Introduction
Deep learning network
Convolutional neural networks
Recurrent neural network
Long short-term memory
Autoencoder
Applications
Transfer learning
Computational biology
Transcriptomics
Splicing analysis
Gene expression
Genomics
Drug discovery
Computer-aided drug design
New drug molecule identification
Protein engineering
Gene expression data analysis
Pharmacodynamics modeling
Medical images
Image segmentation
Image registration
Computer-aided diagnosis
Physical simulation
Electronic health records
Medical informatics
Public health
Deep learning in healthcare: Limitations and challenges
References
Autonomous experimentation in nanotechnology
Introduction
Development of AE capabilities
Automation and robotics
Domain knowledge, discovery systems, and artificial intelligence
Machine learning, optimization, and data analysis
Overview of the development of autonomous experimentation (AE) systems
Case studies of AE in nanotechnology
Carbon nanotube synthesis
Optoelectronic properties of thin films
Synthesis and optimization of nanoparticles
Enhancement observed due to the introduction of automation and autonomy
Platform technologies for AE in nanoscience
Platforms for miniaturized liquid handling
Scanning probes for patterning and interrogating nanoscale materials
Conclusions and future directions
References
Nanomaterials and artificial intelligence in anti-counterfeiting
Introduction
Encryption mechanism of optical security labels
Advanced optical nanomaterials for anti-counterfeiting applications
Photonic crystals
Luminescent materials
Plasmonic materials
Advanced optical anti-counterfeiting labels
Conventional optical anti-counterfeiting labels
Structural color-based anti-counterfeiting labels
Luminescent anti-counterfeiting labels
Plasmonic anti-counterfeiting labels
Physical unclonable function (PUF)-based optical anti-counterfeiting labels
Artificial intelligence-based authentication
Summary and outlook
References
Machine learning data processing as a bridge between microscopy and the brain
Introduction
Machine learning
Identifying active neurons
2D segmentation methods
3D block segmentation methods
3D frame-by-frame segmentation methods
Spike inference
Discussion
References
Index