Heliyon. 2024 Oct 10;10(20):e39185. doi: 10.1016/j.heliyon.2024.e39185. eCollection 2024 Oct 30.

ABSTRACT

Accurate assessment of the three-dimensional (3D) pore characteristics within porous materials and devices holds significant importance. Compared to high-cost experimental approaches, this study introduces an alternative method: utilizing a generative adversarial network (GAN) to reconstruct a 3D pore microstructure. Unlike some existing GAN models that require 3D images as training data, the proposed model only requires a single cross-sectional image for 3D reconstruction. Using porous ceramic electrode materials as a case study, a comparison between the GAN-generated microstructures and those reconstructed through focused ion beam-scanning electron microscopy (FIB-SEM) reveals promising consistency. The GAN-based reconstruction technique demonstrates its effectiveness by successfully characterizing pore attributes in porous ceramics, with measurements of porosity, pore size, and tortuosity factor exhibiting notable agreement with the results obtained from mercury intrusion porosimetry.

PMID:39640653 | PMC:PMC11620251 | DOI:10.1016/j.heliyon.2024.e39185

Heliyon. 2024 Oct 9;10(20):e39038. doi: 10.1016/j.heliyon.2024.e39038. eCollection 2024 Oct 30.

ABSTRACT

Pre-trained chemical language models (CLMs) have attracted increasing attention within the domains of cheminformatics and bioinformatics, inspired by their remarkable success in the natural language processing (NLP) domain such as speech recognition, text analysis, translation, and other objectives associated with language. Furthermore, the vast amount of unlabeled data associated with chemical compounds or molecules has emerged as a crucial research focus, prompting the need for CLMs with reasoning capabilities over such data. Molecular graphs and molecular descriptors are the predominant approaches to representing molecules for property prediction in machine learning (ML). However, Transformer-based LMs have recently emerged as de-facto powerful tools in deep learning (DL), showcasing outstanding performance across various NLP downstream tasks, particularly in text analysis. Within the realm of pre-trained transformer-based LMs such as, BERT (and its variants) and GPT (and its variants) have been extensively explored in the chemical informatics domain. Various learning tasks in cheminformatics such as the text analysis that necessitate handling of chemical SMILES data which contains intricate relations among elements or atoms, have become increasingly prevalent. Whether the objective is predicting molecular reactions or molecular property prediction, there is a growing demand for LMs capable of learning molecular contextual information within SMILES sequences or strings from text inputs (i.e., SMILES). This review provides an overview of the current state-of-the-art of chemical language Transformer-based LMs in chemical informatics for de novo design, and analyses current limitations, challenges, and advantages. Finally, a perspective on future opportunities is provided in this evolving field.

PMID:39640612 | PMC:PMC11620068 | DOI:10.1016/j.heliyon.2024.e39038

Comput Biol Med. 2024 Dec 4;185:109436. doi: 10.1016/j.compbiomed.2024.109436. Online ahead of print.

ABSTRACT

PURPOSE: Accurate and automated early survival prediction is critical for patients with glioblastoma (GBM) as their poor prognosis requires timely treatment decision-making. To address this need, we developed a deep learning (DL)-based end-to-end workflow for GBM overall survival (OS) prediction using pre-resection basic structural multiparametric magnetic resonance images (Bas-mpMRI) with a multi-institutional public dataset and evaluated it with an independent dataset of patients on a prospective institutional clinical trial.

MATERIALS AND METHODS: The proposed end-to-end workflow includes a skull-stripping model, a GBM sub-region segmentation model and an ensemble learning-based OS prediction model. The segmentation model utilizes skull-stripped Bas-mpMRIs to segment three GBM sub-regions. The segmented GBM is fed into the contrastive learning-based OS prediction model to classify the patients into different survival groups. Our datasets include both a multi-institutional public dataset from Medical Image Computing and Computer Assisted Intervention (MICCAI) Brain Tumor Segmentation (BraTS) challenge 2020 with 235 patients, and an institutional dataset from a 5-fraction SRS clinical trial with 19 GBM patients. Each data entry consists of pre-operative Bas-mpMRIs, survival days and patient ages. Basic clinical characteristics are also available for SRS clinical trial data. The multi-institutional public dataset was used for workflow establishing (90% of data) and initial validation (10% of data). The validated workflow was then evaluated on the institutional clinical trial data.

RESULTS: Our proposed OS prediction workflow achieved an area under the curve (AUC) of 0.86 on the public dataset and 0.72 on the institutional clinical trial dataset to classify patients into 2 OS classes as long-survivors (>12 months) and short-survivors (<12 months), despite the large variation in Bas-mpMRI protocols. In addition, as part of the intermediate results, the proposed workflow can also provide detailed GBM sub-regions auto-segmentation with a whole tumor Dice score of 0.91.

CONCLUSION: Our study demonstrates the feasibility of employing this DL-based end-to-end workflow to predict the OS of patients with GBM using only the pre-resection Bas-mpMRIs. This DL-based workflow can be potentially applied to assist timely clinical decision-making.

PMID:39637462 | DOI:10.1016/j.compbiomed.2024.109436

Comput Biol Med. 2024 Dec 4;185:109510. doi: 10.1016/j.compbiomed.2024.109510. Online ahead of print.

ABSTRACT

Proteins are involved in a variety of functions in living organisms. The mutation of amino acid residues in a protein alters its structure, stability, binding, and function, with some mutations leading to diseases. Understanding the influence of mutations on protein structure and function help to gain deep insights on the molecular mechanism of diseases and devising therapeutic strategies. Hence, several generic and disease-specific methods have been proposed to reveal pathogenic effects on mutations. In this review, we focus on the development of prediction methods for identifying disease causing mutations in proteins. We briefly outline the existing databases for disease-causing mutations, followed by a discussion on sequence- and structure-based features used for prediction. Further, we discuss computational tools based on machine learning, deep learning and large language models for detecting disease-causing mutations. Specifically, we emphasize the advances in predicting hotspots and mutations for targets involved in cancer, neurodegenerative and infectious diseases as well as in membrane proteins. The computational resources including databases and algorithms understanding/predicting the effect of mutations will be listed. Moreover, limitations of existing methods and possible improvements will be discussed.

PMID:39637461 | DOI:10.1016/j.compbiomed.2024.109510

Comput Biol Med. 2024 Dec 4;185:109489. doi: 10.1016/j.compbiomed.2024.109489. Online ahead of print.

ABSTRACT

BACKGROUND: A growing body of research is using deep learning to explore the relationship between treatment biomarkers for lung cancer patients and cancer tissue morphology on digitized whole slide images (WSIs) of tumour resections. However, these WSIs typically contain non-cancer tissue, introducing noise during model training. As digital pathology models typically start with splitting WSIs into tiles, we propose a model that can be used to exclude non-cancer tiles from the WSIs of lung squamous cell carcinoma (SqCC) tumours.

METHODS: We obtained 116 WSIs of tumours from 35 different centres from the Cancer Genome Atlas. A pathologist completed or reviewed cancer contours in four regions of interest (ROIs) within each WSIs. We then split the ROIs into tiles labelled with the percentage of cancer tissue within them and trained VGG16 to predict this value, and then we calculated regression error. To measure classification performance and visualize the classification results, we thresholded the predictions and calculated the area under the receiver operating characteristic curve (AUC).

RESULTS: The model's median regression error was 4% with a standard deviation of 35%. At a cancer threshold of 50%, the model had an AUC of 0.83. False positives tended to be in tissues that surround cancer, tiles with <50% cancer, and areas with high immune activity. False negatives tended to be microtomy defects.

CONCLUSIONS: With further validation for each specific research application, the model we describe in this paper could facilitate the development of more effective research pipelines for predicting treatment biomarkers for lung SqCC.

PMID:39637460 | DOI:10.1016/j.compbiomed.2024.109489

Comput Biol Med. 2024 Dec 4;185:109504. doi: 10.1016/j.compbiomed.2024.109504. Online ahead of print.

ABSTRACT

INTRODUCTION: In Parkinson's Disease (PD), despite available treatments focusing on symptom alleviation, the effectiveness of conventional therapies decreases over time. This study aims to enhance the identification of candidates for device-aided therapies (DAT) using artificial intelligence (AI), addressing the need for improved treatment selection in advanced PD stages.

METHODS: This national, multicenter, cross-sectional, observational study involved 1086 PD patients across Spain. Machine learning (ML) algorithms, including CatBoost, support vector machine (SVM), and logistic regression (LR), were evaluated for their ability to identify potential DAT candidates based on clinical and demographic data.

RESULTS: The CatBoost algorithm demonstrated superior performance in identifying DAT candidates, with an area under the curve (AUC) of 0.95, sensitivity of 0.91, and specificity of 0.88. It outperformed other ML models in balanced accuracy and negative predictive value. The model identified 23 key features as predictors for suitability for DAT, highlighting the importance of daily "off" time, doses of oral levodopa/day, and PD duration. Considering the 5-2-1 criteria, the algorithm identified a decision threshold for DAT candidates as > 4 times levodopa tablets taken daily and/or ≥1.8 h in daily "off" time.

CONCLUSION: The study developed a highly discriminative CatBoost model for identifying PD patients candidates for DAT, potentially improving timely and accurate treatment selection. This AI approach offers a promising tool for neurologists, particularly those less experienced with DAT, to optimize referral to Movement Disorder Units.

PMID:39637457 | DOI:10.1016/j.compbiomed.2024.109504

Comput Biol Med. 2024 Dec 4;185:109494. doi: 10.1016/j.compbiomed.2024.109494. Online ahead of print.

ABSTRACT

Cancer, a global health threat, demands effective diagnostic solutions to combat its impact on public health, particularly for breast, colon, and lung cancers. Early and accurate diagnosis is essential for successful treatment, prompting the rise of Computer-Aided Diagnosis Systems as reliable and cost-effective tools. Histopathology, renowned for its precision in cancer imaging, has become pivotal in the diagnostic landscape of breast, colon, and lung cancers. However, while deep learning models have been widely explored in this domain, they often face challenges in generalizing to diverse clinical settings and in efficiently capturing both local and global feature representations, particularly for multi-class tasks. This underscores the need for models that can reduce biases, improve diagnostic accuracy, and minimize error susceptibility in cancer classification tasks. To this end, we introduce ResoMergeNet (RMN), an advanced deep-learning model designed for both multi-class and binary cancer classification using histopathological images of breast, colon, and lung. ResoMergeNet integrates the Resboost mechanism which enhances feature representation, and the ConvmergeNet mechanism which optimizes feature extraction, leading to improved diagnostic accuracy. Comparative evaluations against state-of-the-art models show ResoMergeNet's superior performance. Validated on the LC-25000 and BreakHis (400× and 40× magnifications) datasets, ResoMergeNet demonstrates outstanding performance, achieving perfect scores of 100 % in accuracy, sensitivity, precision, and F1 score for binary classification. For multi-class classification with five classes from the LC25000 dataset, it maintains an impressive 99.96 % across all performance metrics. When applied to the BreakHis dataset, ResoMergeNet achieved 99.87 % accuracy, 99.75 % sensitivity, 99.78 % precision, and 99.77 % F1 score at 400× magnification. At 40× magnification, it still delivered robust results with 98.85 % accuracy, sensitivity, precision, and F1 score. These results emphasize the efficacy of ResoMergeNet, marking a substantial advancement in diagnostic and prognostic systems for breast, colon, and lung cancers. ResoMergeNet's superior diagnostic accuracy can significantly reduce diagnostic errors, minimize human biases, and expedite clinical workflows, making it a valuable tool for enhancing cancer diagnosis and treatment outcomes.

PMID:39637456 | DOI:10.1016/j.compbiomed.2024.109494

Biomed Phys Eng Express. 2024 Dec 5. doi: 10.1088/2057-1976/ad9aab. Online ahead of print.

ABSTRACT

Cardiovascular diseases rank among the leading causes of mortality worldwide and the early identification of diseases is of paramount importance. This work focuses on developing a novel machine learning-based framework for early detection and classification of heart murmurs by analysing phonocardiogram signals. Our heart murmur detection and classification pipeline encompasses three classification settings. We first develop a set of methods based on transfer learning to determine the existence of heart murmurs and categorize them as present, absent, or unknown. If a murmur is present it will be classified as normal or abnormal based on its clinical&#xD;outcome by using 1D convolution and audio spectrogram transformers. Finally, we use Wav2Vec encoder with raw audio data and AdaBoost abstain classifier for heart murmur quality identification. Heart murmurs are categorized based on their specific attributes, including murmur pitch, murmur shape, and murmur timing which are important for diagnosis. Using the PhysioNet 2022 dataset for training and validation, we achieve an 81.08% validation accuracy for murmur presence classification and a 68.23% validation accuracy for clinical outcome classification with 60.52% sensitivity and 74.46% specificity. The suggested approaches provide a promising framework for using phonocardiogram signals for the detection, classification, and quality analysis of heart murmurs. This has significant implications for the diagnosis and treatment of cardiovascular diseases.

PMID:39637440 | DOI:10.1088/2057-1976/ad9aab

PLoS One. 2024 Dec 5;19(12):e0311967. doi: 10.1371/journal.pone.0311967. eCollection 2024.

ABSTRACT

Environmental challenges are rarely confined to national, disciplinary, or linguistic domains. Convergent solutions require international collaboration and equitable access to new technologies and practices. The ability of international, multidisciplinary and multilingual research teams to work effectively can be challenging. A major impediment to innovation in diverse teams often stems from different understandings of the terminology used. These can vary greatly according to the cultural and disciplinary backgrounds of the team members. In this paper we take an empirical approach to examine sources of terminological confusion and their effect in a technically innovative, multidisciplinary, multinational, and multilingual research project, adhering to Open Science principles. We use guided reflection of participant experience in two contrasting teams-one applying Deep Learning (Artificial Intelligence) techniques, the other developing guidance for Open Science practices-to identify and classify the terminological obstacles encountered and reflect on their impact. Several types of terminological incongruities were identified, including fuzziness in language, disciplinary differences and multiple terms for a single meaning. A novel or technical term did not always exist in all domains, or if known, was not fully understood or adopted. Practical matters of international data collection and comparison included an unanticipated need to incorporate different types of data labels from country to country, authority to authority. Sometimes these incongruities could be solved quickly, sometimes they stopped the workflow. Active collaboration and mutual trust across the team enhanced workflows, as incompatibilities were resolved more speedily than otherwise. Based on the research experience described in this paper, we make six recommendations accompanied by suggestions for their implementation to improve the success of similar multinational, multilingual and multidisciplinary projects. These recommendations are conceptual drawing on a singular experience and remain to be sources for discussion and testing by others embarking on their research journey.

PMID:39637194 | DOI:10.1371/journal.pone.0311967

PLoS One. 2024 Dec 5;19(12):e0315127. doi: 10.1371/journal.pone.0315127. eCollection 2024.

ABSTRACT

Convolutional neural network (CNN)-based deep learning (DL) methods have transformed the analysis of geospatial, Earth observation, and geophysical data due to their ability to model spatial context information at multiple scales. Such methods are especially applicable to pixel-level classification or semantic segmentation tasks. A variety of R packages have been developed for processing and analyzing geospatial data. However, there are currently no packages available for implementing geospatial DL in the R language and data science environment. This paper introduces the geodl R package, which supports pixel-level classification applied to a wide range of geospatial or Earth science data that can be represented as multidimensional arrays where each channel or band holds a predictor variable. geodl is built on the torch package, which supports the implementation of DL using the R and C++ languages without the need for installing a Python/PyTorch environment. This greatly simplifies the software environment needed to implement DL in R. Using geodl, geospatial raster-based data with varying numbers of bands, spatial resolutions, and coordinate reference systems are read and processed using the terra package, which makes use of C++ and allows for processing raster grids that are too large to fit into memory. Training loops are implemented with the luz package. The geodl package provides utility functions for creating raster masks or labels from vector-based geospatial data and image chips and associated masks from larger files and extents. It also defines a torch dataset subclass for geospatial data for use with torch dataloaders. UNet-based models are provided with a variety of optional ancillary modules or modifications. Common assessment metrics (i.e., overall accuracy, class-level recalls or producer's accuracies, class-level precisions or user's accuracies, and class-level F1-scores) are implemented along with a modified version of the unified focal loss framework, which allows for defining a variety of loss metrics using one consistent implementation and set of hyperparameters. Users can assess models using standard geospatial and remote sensing metrics and methods and use trained models to predict to large spatial extents. This paper introduces the geodl workflow, design philosophy, and goals for future development.

PMID:39637071 | DOI:10.1371/journal.pone.0315127

PLoS One. 2024 Dec 5;19(12):e0315031. doi: 10.1371/journal.pone.0315031. eCollection 2024.

ABSTRACT

Tomato cultivation is expanding rapidly, but the tomato sector faces significant challenges from various sources, including environmental (abiotic stress) and biological (biotic stress or disease) threats, which adversely impact the crop's growth, reproduction, and overall yield potential. The objective of this work is to build deep learning based lightweight convolutional neural network (CNN) architecture for the real-time classification of biotic stress in tomato plant leaves. This model proposes to address the drawbacks of conventional CNNs, which are resource-intensive and time-consuming, by using optimization methods that reduce processing complexity and enhance classification accuracy. Traditional plant disease classification methods predominantly utilize CNN based deep learning techniques, originally developed for fundamental image classification tasks. It relies on computationally intensive CNNs, hindering real-time application due to long training times. To address this, a lighter CNN framework is proposed to enhance with two key components. Firstly, an Elephant Herding Optimization (EHO) algorithm selects pertinent features for classification tasks. The classification module integrates a Hessian-based Optimal Brain Surgeon (HOBS) approach with a pruned Extreme Learning Machine (ELM), optimizing network parameters while reducing computational complexity. The proposed pruned model gives an accuracy of 95.73%, Cohen's kappa of 0.81%, training time of 2.35sec on Plant Village dataset, comprising 8,000 leaf images across 10 distinct classes of tomato plant, which demonstrates that this framework effectively reduces the model's size of 9.2Mb and parameters by reducing irrelevant connections in the classification layer. The proposed classifier performance was compared to existing deep learning models, the experimental results show that the pruned DenseNet achieves an accuracy of 86.64% with a model size of 10.6 MB, while GhostNet reaches an accuracy of 92.15% at 10.9 MB. CACPNET demonstrates an accuracy of 92.4% with a model size of 18.0 MB. In contrast, the proposed approach significantly outperforms these models in terms of accuracy and processing time.

PMID:39637070 | DOI:10.1371/journal.pone.0315031

PLoS One. 2024 Dec 5;19(12):e0314533. doi: 10.1371/journal.pone.0314533. eCollection 2024.

ABSTRACT

Forensic entomology can help estimate the postmortem interval in criminal investigations. In particular, forensically important fly species that can be found on a body and in its environment at various times after death provide valuable information. However, the current method for identifying fly species is labor intensive, expensive, and may become more serious in view of a shortage of specialists. In this paper, we propose the use of computer vision and deep learning to classify adult flies according to three different families, Calliphoridae, Sarcophagidae, Rhiniidae, and their corresponding genera Chrysomya, Lucilia, Sarcophaga, Rhiniinae, and Stomorhina, which can lead to efficient and accurate estimation of time of death, for example, with the use of camera-equipped drones. The development of such a deep learning model for adult flies may be particularly useful in crisis situations, such as natural disasters and wars, when people disappear. In these cases drones can be used for searching large areas. In this study, two models were evaluated using transfer learning with MobileNetV3-Large and VGG19. Both models achieved a very high accuracy of 99.39% and 99.79%. In terms of inference time, the MobileNetV3-Large model was faster with an average time per step of 1.036 seconds than the VGG19 model, which took 2.066 seconds per step. Overall, the results highlight the potential of deep learning models for the classification of fly species in forensic entomology and search and rescue operations.

PMID:39637032 | DOI:10.1371/journal.pone.0314533

Health Data Sci. 2024 Dec 4;4:0176. doi: 10.34133/hds.0176. eCollection 2024.

ABSTRACT

Background: Missing data in electronic health records (EHRs) presents significant challenges in medical studies. Many methods have been proposed, but uncertainty exists regarding the current state of missing data addressing methods applied for EHR and which strategy performs better within specific contexts. Methods: All studies referencing EHR and missing data methods published from their inception until 2024 March 30 were searched via the MEDLINE, EMBASE, and Digital Bibliography and Library Project databases. The characteristics of the included studies were extracted. We also compared the performance of various methods under different missingness scenarios. Results: After screening, 46 studies published between 2010 and 2024 were included. Three missingness mechanisms were simulated when evaluating the missing data methods: missing completely at random (29/46), missing at random (20/46), and missing not at random (21/46). Multiple imputation by chained equations (MICE) was the most popular statistical method, whereas generative adversarial network-based methods and the k nearest neighbor (KNN) classification were the common deep-learning-based or traditional machine-learning-based methods, respectively. Among the 26 articles comparing the performance among medical statistical and machine learning approaches, traditional machine learning or deep learning methods generally outperformed statistical methods. Med.KNN and context-aware time-series imputation performed better for longitudinal datasets, whereas probabilistic principal component analysis and MICE-based methods were optimal for cross-sectional datasets. Conclusions: Machine learning methods show significant promise for addressing missing data in EHRs. However, no single approach provides a universally generalizable solution. Standardized benchmarking analyses are essential to evaluate these methods across different missingness scenarios.

PMID:39635227 | PMC:PMC11615160 | DOI:10.34133/hds.0176

Nanophotonics. 2023 Oct 2;12(20):3883-3894. doi: 10.1515/nanoph-2023-0298. eCollection 2023 Oct.

ABSTRACT

State-of-the-art deep learning models can converse and interact with humans by understanding their emotions, but the exponential increase in model parameters has triggered an unprecedented demand for fast and low-power computing. Here, we propose a microcomb-enabled integrated optical neural network (MIONN) to perform the intelligent task of human emotion recognition at the speed of light and with low power consumption. Large-scale tensor data can be independently encoded in dozens of frequency channels generated by the on-chip microcomb and computed in parallel when flowing through the microring weight bank. To validate the proposed MIONN, we fabricated proof-of-concept chips and a prototype photonic-electronic artificial intelligence (AI) computing engine with a potential throughput up to 51.2 TOPS (tera-operations per second). We developed automatic feedback control procedures to ensure the stability and 8 bits weighting precision of the MIONN. The MIONN has successfully recognized six basic human emotions, and achieved 78.5 % accuracy on the blind test set. The proposed MIONN provides a high-speed and energy-efficient neuromorphic computing hardware for deep learning models with emotional interaction capabilities.

PMID:39635194 | PMC:PMC11501890 | DOI:10.1515/nanoph-2023-0298

Nanophotonics. 2024 Aug 29;13(22):4159-4168. doi: 10.1515/nanoph-2024-0328. eCollection 2024 Sep.

ABSTRACT

Spectroscopy is a technique that analyzes the interaction between matter and light as a function of wavelength. It is the most convenient method for obtaining qualitative and quantitative information about an unknown sample with reasonable accuracy. However, traditional spectroscopy is reliant on bulky and expensive spectrometers, while emerging applications of portable, low-cost and lightweight sensing and imaging necessitate the development of miniaturized spectrometers. In this study, we have developed a computational spectroscopy method that can provide single-shot operation, sub-nanometer spectral resolution, and direct materials characterization. This method is enabled by a metasurface integrated computational spectrometer and deep learning algorithms. The identification of critical parameters of optical cavities and chemical solutions is demonstrated through the application of the method, with an average spectral reconstruction accuracy of 0.4 nm and an actual measurement error of 0.32 nm. The mean square errors for the characterization of cavity length and solution concentration are 0.53 % and 1.21 %, respectively. Consequently, computational spectroscopy can achieve the same level of spectral accuracy as traditional spectroscopy while providing convenient, rapid material characterization in a variety of scenarios.

PMID:39635447 | PMC:PMC11501049 | DOI:10.1515/nanoph-2024-0328

Nanophotonics. 2021 Oct 14;11(4):779-786. doi: 10.1515/nanoph-2021-0489. eCollection 2022 Jan.

ABSTRACT

Structured light with customized topological patterns inspires diverse classical and quantum investigations underpinned by accurate detection techniques. However, the current detection schemes are limited to vortex beams with a simple phase singularity. The precise recognition of general structured light with multiple singularities remains elusive. Here, we report deep learning (DL) framework that can unveil multi-singularity phase structures in an end-to-end manner, after feeding only two intensity patterns upon beam propagation. By outputting the phase directly, rich and intuitive information of twisted photons is unleashed. The DL toolbox can also acquire phases of Laguerre-Gaussian (LG) modes with a single singularity and other general phase objects likewise. Enabled by this DL platform, a phase-based optical secret sharing (OSS) protocol is proposed, which is based on a more general class of multi-singularity modes than conventional LG beams. The OSS protocol features strong security, wealthy state space, and convenient intensity-based measurements. This study opens new avenues for large-capacity communications, laser mode analysis, microscopy, Bose-Einstein condensates characterization, etc.

PMID:39635381 | PMC:PMC11501744 | DOI:10.1515/nanoph-2021-0489

Nanophotonics. 2023 Aug 11;12(18):3589-3601. doi: 10.1515/nanoph-2023-0291. eCollection 2023 Sep.

ABSTRACT

The selective broadband absorption of solar radiation plays a crucial role in applying solar energy. However, despite being a decade-old technology, the rapid and precise designs of selective absorbers spanning from the solar spectrum to the infrared region remain a significant challenge. This work develops a high-performance design paradigm that combines deep learning and multi-objective double annealing algorithms to optimize multilayer nanostructures for maximizing solar spectral absorption and minimum infrared radiation. Based on deep learning design, we experimentally fabricate the designed absorber and demonstrate its photothermal effect under sunlight. The absorber exhibits exceptional absorption in the solar spectrum (calculated/measured = 0.98/0.94) and low average emissivity in the infrared region (calculated/measured = 0.08/0.19). This absorber has the potential to result in annual energy savings of up to 1743 kW h/m2 in areas with abundant solar radiation resources. Our study opens a powerful design method to study solar-thermal energy harvesting and manipulation, which will facilitate for their broad applications in other engineering applications.

PMID:39635349 | PMC:PMC11502052 | DOI:10.1515/nanoph-2023-0291

Nanophotonics. 2023 Apr 13;12(11):1943-1955. doi: 10.1515/nanoph-2022-0746. eCollection 2023 May.

ABSTRACT

Recent advances in inverse-design approaches for discovering optical structures based on desired functional characteristics have reshaped the landscape of nanophotonic structures, where most studies have focused on how light interacts with nanophotonic structures only. When quantum emitters (QEs), such as atoms, molecules, and quantum dots, are introduced to couple to the nanophotonic structures, the light-matter interactions become much more complicated, forming a rapidly developing field - quantum nanophotonics. Typical quantum functional characteristics depend on the intrinsic properties of the QE and its electromagnetic environment created by the nanophotonic structures, commonly represented by a scalar quantity, local-density-of-states (LDOS). In this work, we introduce a generalized inverse-design framework in quantum nanophotonics by taking LDOS as the bridge to connect the nanophotonic structures and the quantum functional characteristics. We take a simple system consisting of QEs sitting on a single multilayer shell-metal-nanoparticle (SMNP) as an example, apply fully-connected neural networks to model the LDOS of SMNP, inversely design and optimize the geometry of the SMNP based on LDOS, and realize desirable quantum characteristics in two quantum nanophotonic problems: spontaneous emission and entanglement. Our work introduces deep learning to the quantum optics domain for advancing quantum device designs; and provides a new platform for practicing deep learning to design nanophotonic structures for complex problems without a direct link between structures and functional characteristics.

PMID:39635698 | PMC:PMC11501149 | DOI:10.1515/nanoph-2022-0746

Nanophotonics. 2021 Dec 22;11(11):2483-2505. doi: 10.1515/nanoph-2021-0660. eCollection 2022 Jun.

ABSTRACT

Nanophotonics inverse design is a rapidly expanding research field whose goal is to focus users on defining complex, high-level optical functionalities while leveraging machines to search for the required material and geometry configurations in sub-wavelength structures. The journey of inverse design begins with traditional optimization tools such as topology optimization and heuristics methods, including simulated annealing, swarm optimization, and genetic algorithms. Recently, the blossoming of deep learning in various areas of data-driven science and engineering has begun to permeate nanophotonics inverse design intensely. This review discusses state-of-the-art optimizations methods, deep learning, and more recent hybrid techniques, analyzing the advantages, challenges, and perspectives of inverse design both as a science and an engineering.

PMID:39635678 | PMC:PMC11502023 | DOI:10.1515/nanoph-2021-0660

Nanophotonics. 2022 Jan 24;11(11):2507-2529. doi: 10.1515/nanoph-2021-0636. eCollection 2022 Jun.

ABSTRACT

A new type of spectrometer that heavily relies on computational technique to recover spectral information is introduced. They are different from conventional optical spectrometers in many important aspects. Traditional spectrometers offer high spectral resolution and wide spectral range, but they are so bulky and expensive as to be difficult to deploy broadly in the field. Emerging applications in machine sensing and imaging require low-cost miniaturized spectrometers that are specifically designed for certain applications. Computational spectrometers are well suited for these applications. They are generally low in cost and offer single-shot operation, with adequate spectral and spatial resolution. The new type of spectrometer combines recent progress in nanophotonics, advanced signal processing and machine learning. Here we review the recent progress in computational spectrometers, identify key challenges, and note new directions likely to develop in the near future.

PMID:39635673 | PMC:PMC11502016 | DOI:10.1515/nanoph-2021-0636