Hasil untuk "Instruments and machines"

Elena Kakoulli

Edge artificial intelligence (AI) workloads generate bursty, heterogeneous traffic on Networks-on-Chip (NoCs) under tight energy and latency constraints. Hybrid NoCs that overlay electronic meshes with silicon photonic express links can reduce long-path latency via wavelength-division multiplexing, but thermal drift and intermittent optical availability complicate routing. This study introduces a decentralized, photonic-aware controller based on Deep Reinforcement Learning (DRL) with Proximal Policy Optimization (PPO). The policy uses router-local observables—per-port buffer occupancy with short histories, hop distance, a local injection estimate, and a per-cycle optical validity signal—and applies action masking so chosen outputs are always feasible; the controller is co-designed with the router pipeline to retain single-cycle decisions and a modest memory footprint. Cycle-accurate simulations with synthetic traffic and benchmark-derived traces evaluate mean packet latency, throughput, and energy per delivered bit against deterministic, adaptive, and recent DRL baselines; ablation studies isolate the roles of optical validity cues and locality. The results show consistent improvements in congestion-forming regimes and on long electronic paths bridged by photonic links, with robustness across mesh sizes and wavelength concurrency. Overall, the evidence indicates that photonic-aware PPO provides a practical, thermally robust control plane for hybrid NoCs and a scalable routing solution for AI-centric manycore and edge systems.

Payel Dhar, B. Jose Ravindra Raj, Amayappanallur Kannan Dasarathy et al.

ABSTRACT The rapid growth of agri‐food industries has led to an alarming increase in waste generation, posing environmental, economic, and sustainability challenges. This review explores recent advancements in the valorization of agri‐food by‐products into value‐added products through green extraction and biorefinery technologies. It emphasizes the recovery of bioactive compounds such as polyphenols, flavonoids, carotenoids, and dietary fibers from fruit, vegetable, dairy, meat, and seafood wastes, highlighting their potential applications in the food, pharmaceutical, cosmetic, and bioenergy sectors. Emerging eco‐friendly extraction techniques—including supercritical and subcritical fluid extraction, enzyme‐assisted extraction, microwave‐ and ultrasound‐assisted methods, and pulsed electric field processing—offer improved yield, purity, and energy efficiency while reducing ecological impact. Despite technological progress, large‐scale adoption remains constrained by high costs, lack of standardization, and limited industrial integration. Key research gaps include the need for techno‐economic assessments, solvent recovery strategies, and life‐cycle evaluations to ensure process scalability and sustainability. Future research should focus on developing hybrid extraction systems, AI‐driven process optimization, and pilot‐scale biorefineries supported by robust policy frameworks and industry–academia collaboration. Overall, agri‐food waste valorization presents a viable pathway toward achieving environmental sustainability and circular bioeconomy goals, enabling a transition from waste‐intensive practices to resource‐efficient and climate‐resilient production systems.

Ramin Mousa, Saeed Chamani, Mohammad Morsali et al.

Skin cancer (SC) is a life-threatening disease where early diagnosis is critical for effective treatment and survival. While deep learning (DL) has advanced skin cancer diagnosis (SCD), current methods generally yield suboptimal accuracy and efficiency due to challenges in extracting multiscale features from dermoscopic images and optimizing complex model parameters through efficient exploration of the space of hyperparameters. To address this, we propose an approach integrating late Discrete Wavelet Transform (DWT) with pre-trained convolutional neural networks (CNNs) and swarm-based optimization. The late DWT decomposes CNN-extracted feature maps into low- and high-frequency components to improve the detection of subtle lesion patterns, while a self-attention mechanism further refines this by weighing feature importance, focusing on relevant diagnostic information. To refine hyperparameters, three novel swarm-based optimizers – Modified Gorilla Troops Optimizer (MGTO), Improved Gray Wolf Optimization (IGWO), and Fox Optimization (FOX) – are employed searching the space of the hyperparameters to fine-tune the model for superior performance. In comparison to existing methods, experiments on the ISIC-2016 and ISIC-2017 datasets show enhanced classification performance, obtaining at least a 1% accuracy gain. Thus, the suggested framework offers a reliable and effective way to diagnose skin cancer automatically.

Songpon Pumjam, Sarut Panjan, Tarinee Tonggoed et al.

Surface electromyography (sEMG) is widely used for decoding motion intent in prosthetic control and rehabilitation, yet the impact of external load on sEMG-to-kinematics mapping remains insufficiently characterized, particularly for wrist flexion-extension This pilot study investigates wrist angle estimation (0–90°) under four discrete counter-torque levels (0, 25, 50, and 75 N·cm) using a multilayer perceptron neural network (MLPNN) regressor with mean absolute value (MAV) features. Multi-channel sEMG was acquired from three healthy participants while performing isotonic wrist extension (clockwise) and flexion (counterclockwise) in a constrained single-degree-of-freedom setup with potentiometer-based ground truth. Signals were filtered and normalized, and MAV features were extracted using a 200 ms sliding window with a 20 ms step. Across all load levels, the within-subject models achieved very high accuracy (R² = 0.9946–0.9982) with test MSE of 1.23–3.75 deg²; extension yielded lower error than flexion, and the largest error was observed in flexion at 25 N·cm. Because the cohort is small (n = 3), the movement is highly constrained, and subject-independent validation and embedded implementation were not evaluated, these results should be interpreted as a best-case baseline rather than evidence of deployable rehabilitation performance. Future work should test multi-DoF wrist motion, freer movement conditions, richer feature sets, and subject-independent validation.

Ezz El-Din Hemdan, Amged Sayed

In recent years, cutting-edge technologies, such as artificial intelligence (AI), blockchain, and digital twin (DT), have revolutionized the healthcare sector by enhancing public health and treatment quality through precise diagnosis, preventive measures, and real-time care capabilities. Despite these advancements, the massive amount of generated biomedical data puts substantial challenges associated with information security, privacy, and scalability. Applying blockchain in healthcare-based digital twins ensures data integrity, immutability, consistency, and security, making it a critical component in addressing these challenges. Federated learning (FL) has also emerged as a promising AI technique to enhance privacy and enable decentralized data processing. This paper investigates the integration of digital twin concepts with blockchain and FL in the healthcare domain, focusing on their architecture and applications. It also explores platforms and solutions that leverage these technologies for secure and scalable medical implementations. A case study on federated learning for electroencephalogram (EEG) signal classification is presented, demonstrating its potential as a diagnostic tool for brain activity analysis and neurological disorder detection. Finally, we highlight the key challenges, emerging opportunities, and future directions in advancing healthcare digital twins with blockchain and federated learning, paving the way for a more intelligent, secure, and privacy-preserving medical ecosystem.

Andika Fadilla Siagian, Suendri Suendri

Retail businesses, particularly hardware stores, often encounter challenges in order management such as delayed deliveries, inaccurate stock tracking, and limited information transparency factors that hinder operational efficiency and customer satisfaction. This study proposes a web-based order management system utilizing Progressive Web Apps (PWA) technology, developed with the Next.js framework. The Periodic Review System (PRS) method is implemented to calculate reorder points based on actual demand and safety stock levels. System development follows the Waterfall model, with data collected through observation, semi-structured interviews, and literature review. Testing confirms that the application enhances stock accuracy, minimizes delivery delays, supports offline access, and meets SEO performance standards. The implementation significantly improves operational efficiency and holds promise for boosting customer loyalty. The study concludes that PWA-based digital systems are practical, scalable solutions for the MSME sector, with future potential for integration of AI, CRM, and real-time analytics.

Jie Ding, Lingyan Weng, Lili Fan et al.

Tourism demand projection is paramount for both corporate operations and destination management, facilitating tourists in crafting bespoke, multifaceted itineraries and enriching their vacation experiences. This study proposes a multi-layer self attention mechanism recommendation algorithm based on dynamic spatial perception, with the aim of refining the analysis of tourists’ emotional inclinations and providing precise estimates of tourism demand. Initially, the model is constructed upon a foundation of multi-layer attention modules, enabling the semantic discovery of proximate entities to the focal scenic locale and employing attention layers to consolidate akin positions, epitomizing them through contiguous vectors. Subsequently, leveraging tourist preferences, the model forecasts the likelihood of analogous attractions as a cornerstone for the recommendation system. Furthermore, an attention mechanism is employed to refine the spatial layout, utilizing the forecasted passenger flow grid to infer tourism demand across multiple scenic locales in forthcoming periods. Ultimately, through scrutiny of data pertaining to renowned tourist destinations in Beijing, the model exhibits an average MAPE of 8.11%, markedly surpassing benchmarks set by alternative deep learning models, thereby underscoring its precision and efficacy. The spatial layout optimization methodology predicated on a multi-layer attention mechanism propounded herein confers substantive benefits to tourism demand prognostication and recommendation systems, promising to elevate the operational standards and customer contentment within the tourism sector.

Raenald Syaputra, Taghfirul Azhima Yoga Siswa, Wawan Joko Pranoto

Banjir merupakan salah satu bencana alam yang sering terjadi di Indonesia, termasuk di Kota Samarinda dengan 18-33 titik desa terdampak dari tahun 2018-2021. Penggunaan machine learning dalam mengklasifikasi bencana banjir sangat penting untuk memprediksi kejadian di masa mendatang. Beberapa penelitian sebelumnya terkait klasifikasi data banjir dalam 3 tahun terakhir telah dilakukan. Namun, dari beberapa penelitian tersebut memunculkan masalah terkait dengan dataset high dimensional yang dapat menurunkan performa model klasifikasi dan menyebabkan overfitting. Selain itu, masalah lain juga muncul dalam hal imbalance data yang menyebabkan bias terhadap kelas mayoritas dan representasi yang tidak akurat. Oleh karena itu, permasalahan dataset high dimensional dan imbalance data merupakan tantangan spesifik yang harus diatas dalam klasifkasi data banjir Kota Samarinda. Penelitian ini bertujuan mengidentifkasi fitur-fitur yang diperoleh dari seleksi fitur Genetic Algorithm (GA) yang memiliki pengaruh terhadap akurasi klasifikasi data banjir Kota Samarinda menggunakan algoritma Support Vector Machine (SVM), serta meningkatkan akurasi klasifikasi data banjir di Kota Samarinda dengan mengimplementasikan algoritma SVM yang dikombinasikan dengan metode Synthetic Minority Oversampling Technique (SMOTE) untuk oversampling, seleksi fitur dengan GA dan optimasi menggunakan Particle Swarm Optimization (PSO). Teknik validasi yang digunakan adalah 10-fold cross validation dan evaluasi performa menggunakan confusion matrix. Data yang digunakan berasal dari BPBD (Badan Penanggulangan Bencana Daerah) dan BMKG (Badan Meteorologi, Klimatologi, dan Geofisika) Kota Samarinda pada tahun 2021-2023 terdiri dari 11 fitur dan 1.095 record. Hasil penelitian menunjukkan bahwa fitur-fitur penting yang terpilih melalui GA adalah temperatur maksimum, kecepatan angin maksimum, arah angin maksimum, arah angin terbanyak, lamanya penyinaran matahari dan kecepatan angin rata-rata. Dengan kombinasi metode SVM, SMOTE, GA dan PSO, akurasi klasifikasi data banjir mencapai 82,28%. Namun, penelitian ini juga menghadapi tantangan seperti kontradiksi hasil dengan penelitian lain terkait penggunaan SMOTE dan variasi hasil akibat karakteristik dataset serta metode pembagian data yang berbeda. Hasil penelitian ini dapat digunakan oleh pemerintah daerah dan badan penanggulangan bencana daerah Kota Samarinda untuk memprediksi kejadian banjir dengan lebih akurat, serta memungkinkan tindakan pencegahan yang lebih efektif. Penerapan hasil penelitian ini dapat meningkatkan efektivitas dalam mitigasi bencana banjir Kota Samarinda.

Cameron J. Hargreaves, Michael W. Gaultois, Luke M. Daniels et al.

Abstract The application of machine learning models to predict material properties is determined by the availability of high-quality data. We present an expert-curated dataset of lithium ion conductors and associated lithium ion conductivities measured by a.c. impedance spectroscopy. This dataset has 820 entries collected from 214 sources; entries contain a chemical composition, an expert-assigned structural label, and ionic conductivity at a specific temperature (from 5 to 873 °C). There are 403 unique chemical compositions with an associated ionic conductivity near room temperature (15–35 °C). The materials contained in this dataset are placed in the context of compounds reported in the Inorganic Crystal Structure Database with unsupervised machine learning and the Element Movers Distance. This dataset is used to train a CrabNet-based classifier to estimate whether a chemical composition has high or low ionic conductivity. This classifier is a practical tool to aid experimentalists in prioritizing candidates for further investigation as lithium ion conductors.

Koichi Fujiwara

Tianqi WANG, Yingzhou ZHANG, Yunlong DI, Dingwen LI, Linlin ZHU

A surge in the amount of information comes with the rapid development of the technology industry.Across all industries, there is a need to collect and utilize vast amounts of data.While this big data holds immense value, it also poses unprecedented challenges to the field of data security.As relational databases serve as a fundamental storage medium for data, they often contain large-scale data rich in content and privacy.In the event of a data leak, significant losses may occur, highlighting the pressing need to safeguard database ownership and verify data ownership.However, existing database watermarking technologies face an inherent tradeoff between improving watermark embedding capacity and reducing data distortion.To address this issue and enhance watermark robustness, a novel robust database watermarking algorithm based on dynamic difference expansion was introduced.The QR code was employed as the watermark, the SVD decomposition of the low frequency part of the image was utilized after Haar wavelet transform.By extracting specific feature values and using residual feature values as the watermark sequence, it was ensured that the same-length watermark sequence contains more information and the embedded watermark length can be reduced.Furthermore, by combining the adaptive differential evolution algorithm and the minimum difference algorithm, the optimal embedding attribute bits were selected to alleviate the problems of low computational efficiency, high data distortion and poor robustness of traditional difference expansion techniques in embedding watermarks, and to improve the embedding capacity of watermarks while reducing the distortion of data.Experimental results demonstrate that the proposed algorithm achieves a high watermark embedding rate with low data distortion.It is resilient against multiple attacks, exhibiting excellent robustness and strong traceability.Compared to existing algorithms, it offers distinct advantages and holds great potential for broad application in the field of data security.

M. Cerezo, Kunal Sharma, Andrew Arrasmith et al.

Abstract Extracting eigenvalues and eigenvectors of exponentially large matrices will be an important application of near-term quantum computers. The variational quantum eigensolver (VQE) treats the case when the matrix is a Hamiltonian. Here, we address the case when the matrix is a density matrix ρ. We introduce the variational quantum state eigensolver (VQSE), which is analogous to VQE in that it variationally learns the largest eigenvalues of ρ as well as a gate sequence V that prepares the corresponding eigenvectors. VQSE exploits the connection between diagonalization and majorization to define a cost function $$C={{{\rm{Tr}}}}(\tilde{\rho }H)$$ C = Tr ( ρ ̃ H ) where H is a non-degenerate Hamiltonian. Due to Schur-concavity, C is minimized when $$\tilde{\rho }=V\rho {V}^{{\dagger} }$$ ρ ̃ = V ρ V † is diagonal in the eigenbasis of H. VQSE only requires a single copy of ρ (only n qubits) per iteration of the VQSE algorithm, making it amenable for near-term implementation. We heuristically demonstrate two applications of VQSE: (1) Principal component analysis, and (2) Error mitigation.

Sanaa Kaddoura, Ganesh Chandrasekaran, Daniela Elena Popescu et al.

The presence of spam content in social media is tremendously increasing, and therefore the detection of spam has become vital. The spam contents increase as people extensively use social media, i.e., Facebook, Twitter, YouTube, and E-mail. The time spent by people using social media is overgrowing, especially in the time of the pandemic. Users get a lot of text messages through social media, and they cannot recognize the spam content in these messages. Spam messages contain malicious links, apps, fake accounts, fake news, reviews, rumors, etc. To improve social media security, the detection and control of spam text are essential. This paper presents a detailed survey on the latest developments in spam text detection and classification in social media. The various techniques involved in spam detection and classification involving Machine Learning, Deep Learning, and text-based approaches are discussed in this paper. We also present the challenges encountered in the identification of spam with its control mechanisms and datasets used in existing works involving spam detection.

Hongyan Xu, Ayten Yiğiter, Jie Chen

Change point analysis has been useful for practical data analytics. In this paper, we provide a new R package, onlineBcp, based on an online Bayesian change point detection algorithm. This R package conveniently outputs the maximum posterior probabilities of multiple change points, loci of change points, basic statistics for segments separated by identified change points, confidence interval for each unknown segment mean and a plot displaying the segmented data. Practically, missing value pre-treatment of the data, before the change point detection algorithm is implemented, is built in this package. In addition, the Kolmogorov–Smirnov test for checking the normality assumption on each segment, post-change point detection, is included as an option in the package for the ease of data analytic and assumption checking flow. When additional data come in, the package provides a function to combine changes identified based on prior data and changes identified based on additional data and thus provides a fast detection of change points in the data stream when new batches of data are collected.

Mahwish Naz, Jamal Hussain Shah, Muhammad Attique Khan et al.

Provocative heart disease is related to ventricular arrhythmias (VA). Ventricular tachyarrhythmia is an irregular and fast heart rhythm that emerges from inappropriate electrical impulses in the ventricles of the heart. Different types of arrhythmias are associated with different patterns, which can be identified. An electrocardiogram (ECG) is the major analytical tool used to interpret and record ECG signals. ECG signals are nonlinear and difficult to interpret and analyze. We propose a new deep learning approach for the detection of VA. Initially, the ECG signals are transformed into images that have not been done before. Later, these images are normalized and utilized to train the AlexNet, VGG-16 and Inception-v3 deep learning models. Transfer learning is performed to train a model and extract the deep features from different output layers. After that, the features are fused by a concatenation approach, and the best features are selected using a heuristic entropy calculation approach. Finally, supervised learning classifiers are utilized for final feature classification. The results are evaluated on the MIT-BIH dataset and achieved an accuracy of 97.6% (using Cubic Support Vector Machine as a final stage classifier).

Qianyuan Li, Shaoying Tian

With the rapid development of urban economy and the acceleration of urbanization, the demand for urban traffic is increasing rapidly. The single traffic-oriented planning does not take into account the requirements of traffic development on resources and the impact on the environment. The traffic construction of most cities can not fully meet the standard of ecotype. In this paper, the vehicle distribution route optimization problem under multiresource constraints such as vehicle energy capacity and vehicle loading capacity is studied, and the static and dynamic models of distribution route optimization under multiresource constraints are established. In the static distribution route optimization model, the distribution route optimization problem with subloops is solved by modifying the network structure and adding virtual resource points. In the dynamic model, the spatiotemporal network model is used to avoid the generation of subloops, and the description of the vehicle distribution route planning problem is more intuitive and accurate. The model enriches the vehicle distribution route selection scheme at the cost of expanding the model scale. And it can solve the time when the vehicle arrives and leaves the customer point. This paper provides a good countermeasure for solving the environmental and social problems in the transportation system under the condition of resource constraints.

XU Tengteng,HUANG Hengjun

Supervised-Latent Dirichlet Distribution Allocation (s-LDA) model cannot handle the multi-label problem and s-LDA model is not correct distribution in the classification model.The Supervised Labled-LDA(sl-LDA) model is proposed by adding a category label based on the response variable.It analyses s-LDA model and existed problem of topic classification,through verifying the classification accuracy of sl-LDA model,the paper classifies the sl-LDA model and s-LDA model.Experimental results in the Chinese and English news corpus show that English corpus classification performance is improved by about 3.80% and Chinese corpus is improved by about 1.77%.

LIANG Yulong,QU Dan,QIU Zeyu

The performance of speaker aware training method based on i-vector is poor because of using MFCC which has the relative poor robustness as the input feature for the extraction of the i-vector.To solve this problem,an improved i-vector based speaker aware training method is proposed.Firstly,a low dimensional feature extraction method based on SVD is proposed,and then the feature extracted by this method is used to replace the MFCC,which can extract better i-vector.Experimental results show that,in the Vystadial_cz corpus,compared with the DNN-HMM speech recognition system and the original i-vector based speaker aware training method,the recognition performance of this method is increased by 1.62% and 1.52% respectively,in the WSJ corpus,the recognition performance of this method is increased by 3.9% and 1.48% respectively.

Abdolvahhab Fetanat, Gholamreza Shafipour

Power industry restructuring has brought new challenges to the generation unit maintenance scheduling problem. Maintenance scheduling establishes the outage time scheduling of units in a particular time horizon. In the restructured power systems, the decision-making process is decentralized where each generating company (GENCO) tries to maximize its own benefit. Therefore, the principle to draw up the unit maintenance scheduling is different from the traditional centralized power systems. The objective function for GENCOs is to minimize his maintenance investment loss. Therefore, he hopes to put its maintenance on the weeks when the market-clearing price is lowest so that maintenance investment loss descends. This paper addresses the unit maintenance scheduling problem of GENCOs in restructured power systems. The problem is formulated as a mixed integer programming problem, and it is solved by using an optimization method known as biogeography-based optimization (BBO). BBO is simple to implement in practice and requires a reasonably small amount of computing time and a small amount of data communication. BBO has been tested by applying it to a GENCO with three generating units. This model consists of an objective function and related constraints, e.g., maintenance window, generation capacity, load and network flow. The simulation result of this method is compared with a classic method. The outcome is very encouraging and proves that BBO is powerful for minimizing GENCOs’ objective function.

Nancy G. Patterson

Dört yıl önce, sekizinci sınıf Edebi Sanatlar dersime Web teknolojisini dâhil etmeye karar verdiğimde Eylül 1999 , öğrencilerimin elektronik metinden anlam inşa etmek için yeni okuma stratejileri geliştirmiş olabileceklerinden şüphelendim. Ayrıca okuyucu olarak rollerinin ince fakat önemli değişimler geçirdiğinden de kuşkulandım. Bu değişimler, elektronik metinlerle ve hiper metnin yapısıyla doğrudan ilintiliydi. Öğrencilerimin, bir hiper metin okudukları zaman, kendilerinden istenen tercihleri yapma konusunda ilk başta kafaları karışmış gibi görünüyordu fakat bu kafa karışıklığı onları çok kısa bir sürede, benim daha canlı bir okuma modu olarak düşündüğüm şeye, onları aradıkları bilgilere götürecek önemli fikir ve görüş linklerindeki metin yığınlarını okudukları yere ulaştırdı. Hiper metinler, özellikle bilgilendirici hiper metinler, daha önce okunan metinlerden farklı biçimde ve belki de daha etkili biçimde okunacak bir çevreye yerleştirilmişlerdir. Hiper metin, internet üzerinde sık karşılaştığımız bağlantılı elektronik metindir. Bağlantılı sözcük veya resimlere tıkladığımızda, internet üzerindeki bir başka yere ulaşmamız mümkündür. Bu okuma şekli, yeni bir ekran ve genelde tümüyle yeni bir konu sunan bu tıklama eylemi, okuru farklı bir yolla anlam inşa etme yaklaşımına davet eder. İngilizce öğretmenleri olarak bizler öğrencilere, basılı metinle yaptıkları kadar hiper metinle işlem yapmaları ve bu metin biçiminden inşa ettikleri anlam üzerine derince düşünmeleri için yardım etmek zorundayız. Hiper metinleri okumanın öğrenciler için farklı bir deneyim olduğunu kabul etmeliyiz. Hiper metinleri farklı okuyabilme arzusu, biz İngilizce öğretmenleri için garip bir durum olmamalıdır. Menü ve billboardları farklı okuyoruz. Dergi ve romanları farklı okuyoruz. Ders kitaplarını ve bilgisayar kullanım kılavuzlarını farklı okuyoruz. Bunlardaki farklılıkları ve içerdikleri şeyleri öğrencilerimizle tartışmakta zorluk çekmiyoruz. Bu yüzden öğrencilere hiper metinleri okuma, anlama ve yazma biçimleri için seçenekler sunmak mantıklı görünmektedir