Efficient Machine Learning Research Articles

Machine learning attack detection based-on stochastic classifier methods for enhancing of routing security in wireless sensor networks

Wireless Sensor Networks (WSNs) are vulnerable to attacks during data transmission, and many techniques have been proposed to detect and secure routing data. In this paper, we introduce a novel stochastic predictive machine learning approach designed to discern untrustworthy events and unreliable routing attributes, aiming to establish an artificial intelligence-based attack detection system for WSNs. Our methodology leverages real-time analysis of the features of simulated WSN routing data. By integrating Hidden Markov Models (HMM) with Gaussian Mixture Models (GMM), we develop a robust classification framework. This framework effectively identifies outliers, pinpoints malicious network behaviors from their origins, and categorizes them as either trusted or untrusted network activities. In addition, dimensionality reduction techniques are used to improve interpretability, reduce computation and processing time, extract uncorrelated features from network data, and optimize performances. The main advantage of our approach is to establish an efficient stochastic machine learning method capable of analyzing and filtering WSN traffic to prevent suspicious and unsafe data, reduce the large dissimilarity in the collected routing features, and rapidly detect attacks before they occur. In this work, we exploit a well-tuned data set that provides a lot of routing information without losing any data. The experimental results show that the proposed stochastic attack detection system can effectively identify and categorize anomalies in wireless sensor networks with high accuracy. The classification rates of the system were found to be around 83.65%, 84.94% and 94.55%, which is significantly better than the existing classification approaches. Furthermore, the proposed system showed a positive prediction value of 11.84% higher than the existing approaches.

Empowering smart toys and devices with stretchable and vivid triboelectric layers for enhanced interaction and monitoring

The evolving landscape of intelligent devices and toys demands seamless interaction, exceptional performance, and visually captivating features. Triboelectric nanogenerators (TENGs) offer an eco-friendly solution for generating electrical energy in smart toys and devices. This research introduces high-power and stretchable triboelectric layers in various colors tailored for toys and wearable gadgets. Combination of the silicone rubber with the ceramic powders, not only improved mechanical properties, but also provided vibrant colors. This process enhanced the fracture strain and the toughness, and reduced the Young's modulus, so stretchability and overall quality were increased. The best composite, the green layer, demonstrated outstanding attributes such as a power density of 6.5 W/m2, a fracture strain of 385 %, a toughness of 1.86 Mj/m3, and a Young's modulus of 0.20 MPa. The green TENG, a sandwich format with an EGaIn electrode, was used for effective monitoring of human body movements. Additionally, a highly efficient machine learning model has been developed for identifying human motions with a remarkable accuracy of 98 %. The study highlighted the considerable power density, voltage, and current exhibited by the colorful TENGs. By integrating these TENGs into circuitry, an engaging toy was created that produces musical notes and displays corresponding pattern on a computer screen when a colorful array was pressed. These colorful nanogenerators hold significant promise for battery-free toys and wearable electronics.

Tricking AI chips into simulating the human brain: A detailed performance analysis

In recent years, significant strides in Artificial Intelligence (AI) have led to various practical applications, primarily centered around training and deployment of deep neural networks (DNNs). These applications, however, require considerable computational resources, predominantly reliant on modern Graphics-Processing Units (GPUs). Yet, the quest for larger and faster DNNs has spurred the creation of specialized AI chips and efficient Machine-Learning (ML) software tools like TensorFlow and PyTorch have been developed for striking a balance between usability and performance. Simultaneously, the field of computational neuroscience shares a similar quest for increased computational power to simulate more extensive and detailed brain models, while also keeping usability high. Although GPUs have also entered this field, programming complexity remains high, resulting in cumbersome simulations. Inspired by AI progress, we introduce a workflow for easily accelerating brain simulations using TensorFlow and evaluate the performance of various, cutting-edge AI chips – including the Graphcore Intelligence-Processing Unit (IPU), GroqChip, Nvidia GPU with Tensor Cores, and Google Tensor-Processing Unit (TPU) – when simulating a biologically detailed as well as simpler brain models. Our model simulations explore the architectural tradeoffs of a modern-day CPU and these four AI platforms by varying computational density, memory requirements and floating-point numerical accuracy. Results show that the GroqChip achieves the best performance for small networks, yet is unable to simulate large-scale networks. At the scale of mammalian brains, the GPU, IPU and TPU achieve speedups ranging from 29x to 1,208x times over CPU runtimes. Remarkably, the TPU sets a new record for the largest, real-time simulation of the inferior-olivary nucleus in the brain. Reduced-accuracy floating-point implementations make some simulation results unreliable for brain research, notably for the GroqChip. Consequently, this work underscores the potential of ML libraries for accelerating brain simulations as well as the critical role of AI-chip numerical accuracy for biophysically realistic brain models.

Enhancing EEG-based brain-computer interface systems through efficient machine learning classification techniques

Advances in the fields of neuroscience and computer science have greatly enhanced the human brain’s ability to communicate and interact with the surrounding environment. In addition, recent steps in machine learning (ML) have increased the use of electroencephalography (EEG)-based BCIs for artificial intelligence (AI) applications. The prevailing challenge in recording EEG sensor data is that the captured signals are mixed with noise, which makes their effective use difficult. Therefore, strengthening the classification stage becomes extremely important and plays a major role in addressing this problem. In this study, we chose five most widely used classification models that obtained the best results in this field and tested them on two open-source databases. We also focused on improving the hyperparameters of each algorithm to obtain best results. Our results indicate excellent results on the first dataset and acceptable for most models on the second, while RF showed superior performance on both with an accuracy of 100% on the first dataset and 86.47% on the second. This was achieved with the lowest training costs, and better performance compared to previous works we evaluated that used the same databases. These results provide valuable insights and advance the development of brain-computer interface (BCI) technology and design.

An efficient synthetic minority oversampling technique-based ensemble learning model to detect COVID-19 severity

The COVID-19 pandemic has highlighted the importance of accurately predicting disease severity to ensure timely intervention and effective allocation of healthcare resources, which can ultimately improve patient outcomes. This study aims to develop an efficient machine learning (ML) model based on patient demographic and clinical data. It utilizes advanced feature engineering techniques to reduce the dimensionality of dataset and address the issue of highly imbalanced data using synthetic minority oversampling technique (SMOTE). The study employs several ensemble learning models, including XGBoost, Random Forest, AdaBoost, voting ensemble, enhanced-weighted voting ensemble, and stack-based ensembles with support vector machine (SVM) and Gaussian Naïve Bayes as meta-learners, to develop the proposed model. The results indicate that the proposed model outperformed the top-performing models reported in previous studies. It achieved an accuracy of 0.978, sensitivity of 1.0, precision of 0.875, F1-score of 0.934, and receiver operating characteristic area under the curve (ROC-AUC) of 0.965. The study identified several features that significantly correlated with COVID-19 severity, which included respiratory rate (breaths per minute), c-reactive proteins, age, and total leukocyte count (TLC) count. The proposed approach presents a promising method for accurate COVID-19 severity prediction, which may prove valuable in assisting healthcare providers in making informed decisions about patient care.

Fluid identification with Graph Transformer using well logging data

The prediction of fluid through well logging is a cornerstone in guiding exploratory efforts in the energy sector. Comprehending the fluid composition beneath the surface empowers exploration teams to effectively gauge the extent, reserves, and caliber of oil and gas resources. This leads to enhanced strategies in exploration and the judicious use of resources. We introduce an innovative machine learning framework named “Graph Transformer” for predicting fluid. This model melds graph convolutional layers with a Transformer module. It excels in decoding spatial and temporal patterns within well logging data, thus unraveling complex geological dependencies by factoring in the interconnectedness of various data points. Additionally, it features a Positional Encoding module to enhance understanding of sequential data points in terms of depth, thereby overcoming the limitations of sequence independence. The Transformer's Multi-Head Self-Attention mechanism is pivotal in discerning and integrating spatial and temporal interconnections across various depths, elevating its capability to represent geological structures. Initially, the model harnesses key well log data like Density, Acoustic, Gamma-ray, and Compensated Neutron Logs for extracting geological features. These insights are then processed through the Graph Transformer to establish relationship between fluid characteristics and logging parameters. Furthermore, we compare this model with other leading models using precision, recall, and accuracy metrics. Experimental findings affirm the model's high accuracy in predicting fluid within intricate geological settings. Its exceptional adaptability makes it apt for various geological conditions and logging tools. Thus, our Graph Transformer model stands out as a sophisticated, efficient machine learning solution in the realm of well logging fluid prediction, offering geologists and engineers precise tools for exploration and development.

Agricultural land use modeling and climate change adaptation: A reinforcement learning approach

AbstractThis paper provides a novel approach to integrate farmers' behavior in spatially explicit agricultural land use modeling to investigate climate change adaptation strategies. More specifically, we develop and apply a computationally efficient machine learning approach based on reinforcement learning to simulate the adoption of agroforestry practices. Using data from an economic experiment with crop farmers in Southeast Germany, our results show that a change in climate, market, and policy conditions shifts the spatial distribution of the uptake of agroforestry systems. Our modeling approach can be used to advance currently used models for ex ante policy analysis by upscaling existing knowledge about farmers behavioral characteristics and combine it with spatially explicit environmental and farm structural data. The approach presents a potential solution for researchers who aim to upscale information, potentially enriching and complementing existing land use modeling approaches.

An efficient machine learning based CPU scheduler for heterogeneous multicore processors

An efficient machine learning based CPU scheduler for heterogeneous multicore processors

MieAI: a neural network for calculating optical properties of internally mixed aerosol in atmospheric models

Aerosols influence weather and climate by interacting with radiation through absorption and scattering. These effects heavily rely on the optical properties of aerosols, which are mainly governed by attributes such as morphology, size distribution, and chemical composition. These attributes undergo continuous changes due to chemical reactions and aerosol micro-physics, resulting in significant spatio-temporal variations. Most atmospheric models struggle to incorporate this variability because they use pre-calculated tables to handle aerosol optics. This offline approach often leads to substantial errors in estimating the radiative impacts of aerosols along with posing significant computational burdens. To address this challenge, we introduce a computationally efficient and robust machine learning approach called MieAI. It allows for relatively inexpensive calculation of the optical properties of internally mixed aerosols with a log-normal size distribution. Importantly, MieAI fully incorporates the variability in aerosol chemistry and microphysics. Our evaluation of MieAI against traditional Mie calculations, using number concentrations from the ICOsahedral Nonhydrostatic model with Aerosol and Reactive Trace gases (ICON-ART) simulations, demonstrates that MieAI exhibits excellent predictive accuracy for aerosol optical properties. MieAI achieves this with errors well within 10%, and it operates more than 1000 times faster than the benchmark approach of Mie calculations. Due to its generalized nature, the MieAI approach can be implemented in any chemistry transport model which represents aerosol size distribution in the form of log-normally distributed internally mixed modes. This advancement has the potential to replace frequently employed look-up tables and plays a substantial role in the ongoing attempts to reduce uncertainties in estimating aerosol radiative forcing.

Condition Monitoring of a Cartesian Robot with a Mechanically Damaged Gear to Create a Fuzzy Logic Control and Diagnosis Algorithm

The detection of faults during an operational process constitutes a crucial objective within the framework of developing a control system to monitor the structure of industrial mechanisms. Even minor faults can give rise to significant consequences that require swift resolution. This research investigates the impact of overtension in the tooth belt transmission and heating of the screw transmission worm on the vibration signals in a robotic system. Utilizing FFT techniques, distinct frequency characteristics associated with different faults were identified. Overtension in the tooth belt transmission caused localized oscillations, addressed by adjusting the acceleration and deceleration speeds. Heating of the screw transmission worm led to widespread disturbances affecting servo stress and positioning accuracy. A fuzzy logic algorithm based on spectral analysis was proposed for adaptive control, considering the vibration’s frequency and amplitude. The simulation results demonstrated effective damage mitigation, reducing wear on the mechanical parts. The diagnostic approach, relying on limited data, emphasized the feasibility of identifying transmission damage, thereby minimizing maintenance costs. This research contributes a comprehensive and adaptive solution for robotic system diagnostics and control, with the proposed fuzzy logic algorithm showing promise for efficient signal processing and machine learning applications.

Maximizing user experience with LLMOps-driven personalized recommendation systems

The integration of LLMOps into personalized recommendation systems marks a significant advancement in managing LLM-driven applications. This innovation presents both opportunities and challenges for enterprises, requiring specialized teams to navigate the complexity of engineering technology while prioritizing data security and model interpretability. By leveraging LLMOps, enterprises can enhance the efficiency and reliability of large-scale machine learning models, driving personalized recommendations aligned with user preferences. Despite ethical considerations, LLMOps is poised for widespread adoption, promising more efficient and secure machine learning services that elevate user experience and shape the future of personalized recommendation systems.

Interpretable machine learning framework to predict gout associated with dietary fiber and triglyceride-glucose index

BackgroundGout prediction is essential for the development of individualized prevention and treatment plans. Our objective was to develop an efficient and interpretable machine learning (ML) model using the SHapley Additive exPlanation (SHAP) to link dietary fiber and triglyceride-glucose (TyG) index to predict gout.MethodsUsing datasets from the National Health and Nutrition Examination Survey (NHANES) (2005–2018) population to study dietary fiber, the TyG index was used to predict gout. After evaluating the performance of six ML models and selecting the Light Gradient Boosting Machine (LGBM) as the optimal algorithm, we interpret the LGBM model for predicting gout using SHAP and reveal the decision-making process of the model.ResultsAn initial survey of 70,190 participants was conducted, and after a gradual exclusion process, 12,645 cases were finally included in the study. Selection of the best performing LGBM model for prediction of gout associated with dietary fiber and TyG index (Area under the ROC curve (AUC): 0.823, 95% confidence interval (CI): 0.798–0.848, Accuracy: 95.3%, Brier score: 0.077). The feature importance of SHAP values indicated that age was the most important feature affecting the model output, followed by uric acid (UA). The SHAP values showed that lower dietary fiber values had a more pronounced effect on the positive prediction of the model, while higher values of the TyG index had a more pronounced effect on the positive prediction of the model.ConclusionThe interpretable LGBM model associated with dietary fiber and TyG index showed high accuracy, efficiency, and robustness in predicting gout. Increasing dietary fiber intake and lowering the TyG index are beneficial in reducing the potential risk of gout.

Efficient Machine Learning Techniques to Classifying Cardiovascular Disease and Improve Prediction Analysis

Cardiovascular Disease (CVD) account for a large portion of the global health burden and are one of the main causes of decease worldwide. In the classification and forecasting of CVDs, Machine Learning (ML) techniques have demonstrated encouraging outcomes. In this research report, a comparative analysis of classification and prediction models for CVD is presented, including both linear and ensemble ML approaches. The paper compares ensemble models like Catboost, Histogram Gradient Boosting Machine (HGBM), and Extra Trees against linear models like Gaussian Nave Bayes, SVM, and KNN. The objective is to identify the most effective CVD prediction model by assessing its performance through accuracy, precision, sensitivity and F1 score as key evaluation metrics. Moreover, results show that ensemble models outperform linear models using advanced techniques such as boosting and histogram-based algorithms. The results underscore the critical role ensemble models play in accurately diagnosing and predicting cardiovascular disease and provide important new information to researchers and healthcare providers. Using these models has the potential to significantly improve patient outcomes and health management by enabling early detection and intervention

Analysis of Soil Health Parameters to Identify Important Soil Nutrients and Weights Using Feature Engineering for Multiple Agri-Advices

Introduction:: Identification of important soil nutrients is a very important task for precision farming and developing efficient machine learning models. Method:: The existing work shows that the patent is filed and published on a method and device for assessment of soil health parameters and recommendation of fertilizers. The existing work is done for one advice at a time not for several advices. Multiple advices that are taken into account for the task are appropriate crops, organic fertilizer, and combination 1 and combination 2 of fertilizers. objective: Apply feature selection techniques based on Chi-Square, ANOVA and Mutual Information Gain scoring functions such as Select K Best and Select Percentile for multiple agri-advice dataset of Pune District regions to identify important soil health features to reduce the complexity of classification models and in turn reduce space and the computational time of different classification models. Result:: This paper presented results of feature selection techniques based on Chi-Square, ANOVA and Mutual Information Gain scoring functions such as Select K Best and Select Percentile for multiple agri-advice datasets of Pune District regions to identify important soil health features. Conclusion:: As per Chi-Square, ANOVA and Mutual Information scoring functions with Select K Best and Select Percentile techniques ‘Mn’ was the most important parameter and Cu’ and ‘B’ were the least important parameters among all 11 parameters common in 4 agriculture advices. Whereas Ph, K, Fe, 'OC', 'N', 'S', 'Mn', and 'P' will be used for future research work on the development of an efficient classification algorithm for multi-advice generators.

A Method of Skin Disease Detection Using Image Processing and Machine Learning

Skin conditions are prevalent health issues globally. The dangers posed by these infections are unseen, leading to physical discomfort and mental anguish. In severe cases, skin diseases can even progress to skin cancer. Consequently, the identification of skin diseases from clinical images remains a significant challenge in medical image analysis. Furthermore, manual diagnosis by medical professionals is time-consuming and subjective. Therefore, there is a need for automated skin disease prediction to expedite treatment planning for both patients and dermatologists. These studies presents a digital hair removal method and de-blurs or denoise the images. To extract essential patterns from the skin imagesand statistical features techniques are utilized. Two efficient machine learning algorithms, namely Support Vector Machine (SVM), and Convolutional Neural Network (CNN) classifiers, are employed with the extracted features to accurately classify skin images as melanoma (MEL), melanocytic nevus (NV), basal cell carcinoma (BCC), actinic keratosis (AK), benign keratosis (BKL), dermatofibroma (DF), vascular lesion (VASC), and Squamous cell carcinoma (SCC). The models are validated using the ISIC 2019 challenge and HAM10000 datasets, with SVM demonstrating slightly superior performance compared to the other classifiers. Furthermore, a comparison with state-of-the-art methods is conducted to evaluate the effectiveness of the proposed approach.

Controlling chaos using edge computing hardware

Machine learning provides a data-driven approach for creating a digital twin of a system – a digital model used to predict the system behavior. Having an accurate digital twin can drive many applications, such as controlling autonomous systems. Often, the size, weight, and power consumption of the digital twin or related controller must be minimized, ideally realized on embedded computing hardware that can operate without a cloud-computing connection. Here, we show that a nonlinear controller based on next-generation reservoir computing can tackle a difficult control problem: controlling a chaotic system to an arbitrary time-dependent state. The model is accurate, yet it is small enough to be evaluated on a field-programmable gate array typically found in embedded devices. Furthermore, the model only requires 25.0 ±7.0\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$\\pm 7.0$$\\end{document} nJ per evaluation, well below other algorithms, even without systematic power optimization. Our work represents the first step in deploying efficient machine learning algorithms to the computing “edge.”

Diagnosing drivers of PM2.5 simulation biases in China from meteorology, chemical composition, and emission sources using an efficient machine learning method

Abstract. Chemical transport models (CTMs) are widely used for air pollution modeling, which suffer from significant biases due to uncertainties in simplified parameterization, meteorological fields, and emission inventories. Accurate diagnosis of simulation biases is critical for the improvement of models, interpretation of results, and management of air quality, especially for the simulation of fine particulate matter (PM2.5). In this study, an efficient method with high speed and a low computational resource requirement based on the tree-based machine learning (ML) method, the light gradient boosting machine (LightGBM), was designed to diagnose CTM simulation biases. The drivers of the Community Multiscale Air Quality (CMAQ) model biases are compared to observations obtained by simulating PM2.5 concentrations from the perspectives of meteorology, chemical composition, and emission sources. The source-oriented CMAQ was used to diagnose the influences of different emission sources on PM2.5 biases. The model can capture the complex relationship between input variables and simulation bias well; meteorology, PM2.5 components, and source sectors can partially explain the simulation bias. The CMAQ model underestimates PM2.5 by −19.25 to −2.66 µg m−3 in 2019, especially in winter and spring and during high-PM2.5 events. Secondary organic components showed the largest contribution to the PM2.5 simulation bias for different regions and seasons (13.8 %–22.6 %) of all components. Relative humidity, cloud cover, and soil surface moisture were the main meteorological factors contributing to PM2.5 bias in the North China Plain, Pearl River Delta, and northwestern China, respectively. Primary and secondary inorganic components from residential sources showed the two largest contributions to this bias (12.05 % and 12.78 %), implying large uncertainties in this sector. The ML-based methods provide valuable complements to traditional-mechanism-based methods for model improvement, with high efficiency and low reliance on prior information.

Efficient Machine Learning and Deep Learning Techniques for Detection of Breast Cancer Tumor

The detection of cancer tumors is an essential component that has important consequences for the speedy involvement of medical professionals and the enhancement of patient outcomes. This review paper presents a complete study of the current body of research and methodology, as well as an in-depth assessment of the use of machine learning (ML) and deep learning (DL) in the detection of cancer tumors. In addition, the article gives a full analysis of the approaches involved. Machine learning and deep learning, which effectively handle ambiguity in the identification of malignant tumors, provide an alternative method for dealing with the complexity of brain tissue. This method is offered by a combination of machine learning and deep learning. The first part of the review draws attention to the significance of making an accurate diagnosis of breast cancer, highlights the limits of traditional diagnostic methods, and investigates the cutting-edge area of medical imaging technology. After that, it investigates the fundamentals of ML and DL and how they might be used to deal with the challenges that are inherent in the interpretation of complicated imaging data. In addition, the paper explores the ways in which models enhance the processes of feature extraction, picture segmentation, and classification in breast tumor detection systems.

Application of a data-driven approach for maximum fatigue damage prediction of an unbonded flexible riser

The fatigue damage prediction is a challenging issue in the design and maintenance phases of unbonded flexible pipes. This arises from the necessity of conducting numerous numerical analyses, encompassing both global performance and local structural assessments, to determine the long-term fatigue damage. To handle the computational-demanding problem for fatigue damage estimation, this work applies an efficient machine learning approach known as the AK-MDAmax approach for maximum fatigue damage estimation of unbonded flexible risers. Kriging surrogate models are employed to predict the short-term distribution of fatigue damage over the sea states. The maximum long-term fatigue damage of the flexible risers is then estimated using the AK-MDAmax approach. The verification process of the AK-MDAmax model unfolds incrementally, testing its feasibility and stability across four different scenarios with varying numbers of considered locations, ranging from 4 to 512. Comparative analyses against numerical results reveal that the AK-MDAmax approach efficiently and accurately estimates maximum long-term fatigue damage across different locations. Remarkably, a 25-fold increase in efficiency is achieved with an error margin of less than 1%. This study demonstrates the substantial potential of the proposed machine learning methodology, offering designers an efficient tool for optimizing flexible risers and mitigating computational costs and time constraints.

Leveraging YOLO Algorithm and PaddleOCR in Machine Learning Applications

Abstract: This research is about implementation of the YOLO algorithm in PaddleOCR and several aspect of machine learning. The mention discusses this technology integration and way they have worked to accomplish the tasks and the expected use on the real world scenarios. The paper accomplishes this by extensively analysing the literature and doing deliberate experimentation. Insights that will give on algorithm effectiveness and challenges are also captured in this paper. Contemporary computer vision systems that leverage efficient machine learning approaches like YOLO (You Only Look Once) and PaddleOCR have expanded in almost every industrial sector. The paper is concerned with the integration of these algorithms in the wide programs and consequential effect on the practical field. A systemic reading of up to date literature and experimental analysis is done by this paper to bring out this vital aspect of their usage, the challenges as well as their prospects in the future