Powerful Learning Research Articles

Robust active learning framework for structural reliability analysis using uncertainty quantification and flexible meta-model

Achieving highly accurate results in assessing structural reliability with reduced time complexity is an active and challenging engineering topic, especially when working with complex real-world structures. One of the most prominent methods addressing this issue is AK-MCS, which adaptively updates the Kriging-based surrogate model for forecasting structures’ behavior. While the Kriging model’s performance may be limited compared to multiple new machine learning models, it remains dominant in reliability analysis thanks to its distinctive ability to assess uncertainty along with prediction results. This ability is necessary to determine the potential candidate sample in the active learning procedure. This study proposes a novel method, dubbed AUQ-meta, that allows us to leverage the recent powerful machine learning models while still being able to produce uncertainty quantification. The AUQ-meta method consists of three main components: a machine learning-based predictor, a quantile regression module for uncertainty assessment, and an adaptive reliability analysis process. The effectiveness and efficiency of the proposed method are consistently validated through six examples with high dimensionality, nonlinearity, and time-varying elements. In addition, comparison and sensitivity studies are carried out, providing further insights into the AUQ-meta method’s operational mechanism.

Position adaptive residual block and knowledge complement strategy for point cloud analysis

Due to the sparsity, irregularity and disorder of the point cloud, the tasks related to it are full of challenges. Exploring local geometric patterns and multi-scale features is effective for point cloud understanding, and promising results have been achieved. In this paper, we present a Position Adaptive Residual Block, namely PARB, for the first time. It can carry out powerful geometric signal description and feature learning. Starting from this module, we propose two extensions. First, a Position Adaptive Residual Network, called PARNet, is derived by utilizing PARB. Second, PARB can be regarded as a plug-and-play module embedded in MLP-based networks, which can remarkably enhance the performance of the backbone. We also introduce an efficient Knowledge Complement Strategy, which is part of the PARNet architecture, to make the framework perform better. Extensive experimental results on challenging benchmarks demonstrate that our PARNet delivers the new state-of-the-art on ShapeNet-Part and achieves competitive performance on ModelNet40.

Data-driven intensive care: alack of comprehensive datasets

Intensive care units provide adata-rich environment with the potential to generate datasets in the realm of big data, which could be utilized to train powerful machine learning (ML) models. However, the currently available datasets are too small and exhibit too little diversity due to their limitation to individual hospitals. This lack of extensive and varied datasets is aprimary reason for the limited generalizability and resulting low clinical utility of current ML models. Often, these models are based on data from single centers and suffer from poor external validity. There is an urgent need for the development of large-scale, multicentric, and multinational datasets. Ensuring data protection and minimizing re-identification risks pose central challenges in this process. The "Amsterdam University Medical Center database (AmsterdamUMCdb)" and the "Salzburg Intensive Care database (SICdb)" demonstrate that open access datasets are possible in Europe while complying with the data protection regulations of the General Data Protection Regulation (GDPR). Another challenge in building intensive care datasets is the absence of semantic definitions in the source data and the heterogeneity of data formats. Establishing binding industry standards for the semantic definition is crucial to ensure seamless semantic interoperability between datasets.

Supervised Contrastive Discrete Hashing for cross-modal retrieval

Cross-modal hashing aims at mapping data points from varied modalities to a shared Hamming space for providing swift and adaptive search services. The recent progress in contrastive learning has significantly enhanced the retrieval performance of cross-modal hashing, owing to its strong ability in powerful representations learning. However, most of them face two primary challenges: (1) they select negative samples for anchors randomly in each mini-batch, which may bring false negative samples that destroy the intrinsic semantic similarity in the learned hash codes; (2) the continuous relaxation strategy is commonly used to generate discrete hash codes, which results in quantization errors and consequently impairing the quality of hash codes. To alleviate these issues, we propose a deep supervised cross-modal hashing scheme, termed Supervised Contrastive Discrete Hashing (SCDH). Specifically, we extend the unsupervised contrastive method to the supervised setting to construct the positive and negative data points with class labels, avoiding the performance decreasing caused by false negative data. Besides, multiple positive and negative samples for each anchor can be chosen from the generated global dictionary bank by class labels, which can better capture the global data structure. Moreover, discrete hash codes can be generated directly, enhancing the feature representational capacity of hash codes. Comprehensive experiments conducted on two public datasets affirm the superior performance of the proposed SCDH over several other cutting-edge methods.

AI-powered malware detection with Differential Privacy for zero trust security in Internet of Things networks

The widespread usage of Android-powered devices in the Internet of Things (IoT) makes them susceptible to evolving cybersecurity threats. Most healthcare devices in IoT networks, such as smart watches, smart thermometers, biosensors, and more, are powered by the Android operating system, where preserving the privacy of user-sensitive data is of utmost importance. Detecting Android malware is thus vital for protecting sensitive information and ensuring the reliability of IoT networks. This article focuses on AI-enabled Android malware detection for improving zero trust security in IoT networks, which requires Android applications to be verified and authenticated before providing access to network resources. The zero trust security model requires strict identity verification for every entity trying to access resources on a private network, regardless of whether they are inside or outside the network perimeter. Our proposed solution, DP-RFECV-FNN, an innovative approach to Android malware detection that employs Differential Privacy (DP) within a Feedforward Neural Network (FNN) designed for IoT networks under the zero trust model. By integrating DP, we ensure the confidentiality of data during the detection process, setting a new standard for privacy in cybersecurity solutions. By combining the strengths of DP and zero trust security with the powerful learning capacity of the FNN, DP-RFECV-FNN demonstrates the ability to identify both known and novel malware types and achieves higher accuracy while maintaining strict privacy controls compared with recent papers. DP-RFECV-FNN achieves an accuracy ranging from 97.78% to 99.21% while utilizing static features and 93.49% to 94.36% for dynamic features of Android applications to detect whether it is malware or benign. These results are achieved under varying privacy budgets, ranging from ϵ=0.1 to ϵ=1.0. Furthermore, our proposed feature selection pipeline enables us to outperform the state-of-the-art by significantly reducing the number of selected features and training time while improving accuracy. To the best of our knowledge, this is the first work to categorize Android malware based on both static and dynamic features through a privacy-preserving neural network model.

Intelligent Low-Consumption Optimization Strategies: Economic Operation of Hydropower Stations Based on Improved LSTM and Random Forest Machine Learning Algorithm

The economic operation of hydropower stations has the potential to increase water use efficiency. However, there are some challenges, such as the fixed and unchangeable flow characteristic curve of the hydraulic turbines, and the large number of variables in optimal load distribution, which limit the progress of research. In this paper, we propose a new optimal method of the economic operation of hydropower stations based on improved Long Short-Term Memory neural network (I-LSTM) and Random Forest (RF) algorithm. Firstly, in order to accurately estimate the water consumption, the LSTM model’s hyperparameters are optimized using improved particle swarm optimization, and the I-LSTM method is proposed to fit the flow characteristic curve of the hydraulic turbines. Secondly, the Random Forest machine learning algorithm is introduced to establish a load-distribution model with its powerful feature extraction and learning ability. To improve the accuracy of the load-distribution model, we use the K-means algorithm to cluster the historical data and optimize the parameters of the Random Forest model. A Hydropower Station in China is selected for a case study. It is shown that (1) the I-LSTM method fits the operating characteristics under various working conditions and actual operating characteristics of hydraulic turbines, ensuring that they are closest to the actual operating state; (2) the I-LSTM method is compared with Support Vector Machine (SVM), Extreme Learning Machine (ELM) and Long Short-Term Memory neural network (LSTM). The prediction results of SVM have a large error, but compared with ELM and LSTM, MSE is reduced by about 46% and 38% respectively. MAE is reduced by about 25% and 21%, respectively. RMSE is reduced by about 27% and 24%, respectively; (3) the RF algorithm performs better than the traditional dynamic programming algorithm in load distribution. With the passage of time and the increase in training samples, the prediction accuracy of the Random Forest model has steadily improved, which helps to achieve optimal operation of the units, reducing their average total water consumption by 1.24%. This study provides strong support for the application of intelligent low-consumption optimization strategies in hydropower fields, which can bring higher economic benefits and resource savings to renewable energy production.

Integrated multi-omics analysis and machine learning developed a prognostic model based on mitochondrial function in a large multicenter cohort for Gastric Cancer

BackgroundGastric cancer (GC) is a common and aggressive type of cancer worldwide. Despite recent advancements in its treatment, the prognosis for patients with GC remains poor. Understanding the mechanisms of cell death in GC, particularly those related to mitochondrial function, is crucial for its development and progression. However, more research is needed to investigate the significance of the interaction between mitochondrial function and GC cell death.MethodsWe employed a robust computational framework to investigate the role of mitochondria-associated proteins in the progression of GC in a cohort of 1,199 GC patients. Ten machine learning algorithms were utilized and combined into 101 unique combinations. Ultimately, we developed a Mitochondrial-related-Score (MitoScore) using the machine learning model that exhibited the best performance. We observed the upregulation of LEMT2 and further explored its function in tumor progression. Mitochondrial functions were assessed by measuring mitochondrial ATP, mitochondrial membrane potential, and levels of lactate, pyruvate, and glucose.ResultsMitoScore showed significant correlations with GC immune and metabolic functions. The higher MitoScore subgroup exhibited enriched metabolic pathways and higher immune activity. Overexpression of LETM2 (leucine zipper and EF-hand containing transmembrane protein 2) significantly enhanced tumor proliferation and metastasis. LETM2 plays a role in promoting GC cell proliferation by activating the mTOR pathway, maintaining mitochondrial homeostasis, and promoting glycolysis.ConclusionThe powerful machine learning framework highlights the significant potential of MitoScore in providing valuable insights and accurate assessments for individuals with GC. This study also enhances our understanding of LETM2 as an oncogene signature in GC. LETM2 may promote tumor progression by maintaining mitochondrial health and activating glycolysis, offering potential targets for diagnosis, treatment, and prognosis of GC.

ANTIBIOTICS RESISTANCE PREDICTION USING MACHINE LEARNING ALGORITHMS

Opioid Risk Antibiotics Resistance abuse, overdose, and drug addiction have become public health concerns as a result of the notable increase in the use of both prescribed and over-the-counter medications in the United States. The growing incidence of opioid usage presents a complex public health issue, one of which is the possible influence on antibiotic resistance. The existing system predicts antibiotic resistance profiles in bacterial infections among hospitalized patients using machine learning (ML) algorithms and electronic medical records (EMRs). However, it ignores the impact of opioid use, a major public health issue that may be connected to antibiotic resistance. The complex association between opioid-related behaviors and antibiotic resistance may not be adequately captured by powerful machine learning algorithms, despite their use. The larger context of antibiotic resistance patterns is limited by the lack of linkage with data on opioid intake. The proposed abstract aims to improve predictive accuracy and understanding of the complex factors influencing antibiotic resistance emergence by integrating support vector machines (SVM) and linear regression to thoroughly analyse the relationship between opioid risk and antibiotic resistance. This approach addresses these limitations. We suggested combining linear regression and support vector machines (SVM) to better understand the complex connection between opioid risk and antibiotic resistance. In order to pinpoint the main causes of increased risk, trends in opioid consumption data are analyzed using the SVM model. In parallel, the relationship between opioid abuse and the rise in antibiotic resistance is examined using linear regression. Our goal is to have a thorough grasp of the complex relationship between opioid-related behaviors and the emergence of antibiotic resistance by combining these two analytical approaches. This project is designed under NetBeans with java as front end. Weka tool for machine learning analysis Keywords: Antibiotic Resistance, Machine Learning, Bacterial Infections, Global Crisis, Healthcare.

WELL: Applying bug detectors to bug localization via weakly supervised learning

AbstractBug localization, which is used to help programmers identify the location of bugs in source code, is an essential task in software development. Researchers have already made efforts to harness the powerful deep learning (DL) techniques to automate it. However, training bug localization model is usually challenging because it requires a large quantity of data labeled with the bug's exact location, which is difficult and time‐consuming to collect. By contrast, obtaining bug detection data with binary labels of whether there is a bug in the source code is much simpler. This paper proposes a WEakly supervised bug LocaLization (WELL) method, which only uses the bug detection data with binary labels to train a bug localization model. With CodeBERT finetuned on the buggy‐or‐not binary labeled data, WELL can address bug localization in a weakly supervised manner. The evaluations on three method‐level synthetic datasets and one file‐level real‐world dataset show that WELL is significantly better than the existing state‐of‐the‐art model in typical bug localization tasks such as variable misuse and other bugs.

Twin support vector machines based on chaotic mapping dung beetle optimization algorithm

Abstract Twin Support Vector Machine (TSVM) is a powerful machine learning method that is usually used to solve binary classification problems. But although the classification speed and performance of TSVM is better than that of primitive support vector machine, TSVM still faces the problem of difficult parameter selection; therefore, to overcome the problem of parameter selection of TSVM, this paper proposes a Chaotic Mapping Dung Beetle Optimization Algorithm-based Twin Support Vector Machine (CMDBO-TSVM) for automatic parameter selection. Due to the uncertainty of the random initialization population of the original Dung Beetle Optimization Algorithm, this paper additionally adds chaotic mapping initialization to improve the Dung Beetle Optimization Algorithm. Experiments on the dataset through this paper show that the classification accuracy of the CMDBO-TSVM has a better performance.

Carbon Emission Forecast Based on Multilayer Perceptron Network and STIRPAT Model

INTRODUCTION: It is of great research significance to explore whether China can achieve the "two-carbon target" on time. The MLP model combines nonlinear modeling principles with other techniques, possessing powerful adaptive learning capabilities, and providing a viable solution for carbon emission prediction.
 OBJECTIVES: This study models and forecasts carbon emissions in Jiangsu Province, one of China's largest industrial provinces, aiming to forecast whether Jiangsu province will achieve the two-carbon target on time plan and provide feasible pathways and theoretical foundations for achieving dual carbon goals.
 METHODS: Based on the analysis of the contributions of relevant indicators using the Grey Relational Analysis method, a comprehensive approach integrating the STIRPAT model, Logistic model, and ARIMA model is adopted. Ultimately, an MLP prediction model for carbon emission variations is established. Using this model, simulations are conducted to analyze the carbon emission levels in Jiangsu Province under different scenarios from 2021 to 2060.
 RESULTS: The time to reach carbon peak and the likelihood of achieving carbon neutrality vary under three scenarios. Under the natural scenario of no human intervention, achieving carbon neutrality is not feasible. While under human-made intervention scenarios including baseline and intervention scenarios, Jiangsu Province is projected to achieve the carbon neutrality target as scheduled, attaining the peak carbon goal, however, proves challenging to realize by the year 2030.
 CONCLUSION: The MLP model exhibits high accuracy in predicting carbon emissions. To expedite the realization of dual carbon goals, proactive government intervention is necessary.

A novel fusion Support Vector Machine integrating weak and sphere models for classification challenges with massive data

The unprecedented growth in data generation has necessitated the adoption of advanced analytical techniques. Support Vector Machine (SVM) is a powerful machine learning tool that has proven invaluable in classifying observations through optimal hyperplane in higher dimensions. Despite their widespread use, SVM models encounter substantial challenges during the learning phase with massive datasets, necessitating strategic modifications. This paper introduces a novel fusion methodology incorporating weak and sphere support vector machine to address classification challenges with massive datasets. Comparative analyses across diverse simulated and benchmark real datasets underscore the efficacy of the proposed methodologies, exhibiting sustained predictive performance. The remarkable efficiency gain is noteworthy, as the learning phase requires only 10% of the computation time compared to conventional SVM approaches.

Self-supervised graph autoencoder with redundancy reduction for community detection

Community detection is a significant research topic in network science, which has been revisited with graph neural networks. As a powerful graph representation learning model, graph autoencoder (GAE) is commonly used for unsupervised community detection. However, most GAE-based methods ignore multi-scale features of encoding layers, which inherently provide useful information for community detection. Moreover, these methods fail to simultaneously improve the representation learning process and clustering performance through a unified objective function. To address these issues, we propose a self-supervised graph autoencoder model with redundancy reduction for community detection. Firstly, we introduce a multi-scale module based on GAE to obtain discriminative node representations from different encoding layers. In particular, a redundancy reduction strategy is employed to eliminate redundancy information in the latent embedding space. Then, a node clustering module is used to obtain initial community labels. To fully utilize the multi-scale features to further refine clustering performance, a self-supervised module is designed to utilize current clustering labels to supervise the node representation learning process, thus constructing an end-to-end model for community detection. Finally, we validate the effectiveness of the proposed method on real-world networks. Experimental results demonstrate that our method outperforms several state-of-the-art methods in community detection.

AirMouse and AirWrite Intuitive Digital Interaction

The primary purpose of this project is to redefine the way people will interact with their systems more easily, and this will be achieved by eliminating the touchpads and keyboards to connect to the computer. This cutting-edge system eliminates the need for a regular mouse or a keyboard, allowing for easier device user-interface adaptation. The implementation of the latest gesture recognition technology lets the system not only pick up and decipher the movements of users' hands but also translate them into wider operations that are similar to those of a mouse and keyboard. Central to the design of the metaphoric system is the "virtual mouse," which enables the cursor movement and the performance of various mouse buttons without touching anything physically. So, it is a virtual mouse with an on-the-screen cursor that reacts to hand gestures in real-time. The system utilizes certain hand gestures enough to enable the effortless execution of actions commonly associated with a traditional mouse including left-clicking, right-clicking, and dragging across the screen. The virtual mouse feature is enhanced by the air writing technique, commonly acknowledged as a milestone in input development. This capability precisely reads and decodes the movements developed in the space by the fingertips while conveying them to the screen as written words. The air writing system would predict the characters from the user's gestures, creating a fresh, secure, and user-friendly method that favors the option of air typing rather than relying on a physical keyboard. This project brings innovation in human-computer interaction when implemented using powerful computer vision and machine learning techniques and promotes accessibility and usability of the system where the traditional inputs become inapplicable. On the other hand, its feasible justifications range from medical uses to pioneering in virtual reality, gaming, and education that drives brilliantly the new options in interaction across various fields.

Predicting Multiple Diseases with Machine Learning and Streamlit: Enhancing Healthcare Digitally

Machine learning (ML) and streamlit technologies have emerged as strong tools in the field of healthcare, with the ability for predicting and diagnosing many diseases with impressive precision and speed. This research study examines the transformational potential of ML algorithms combined with streamlit for disease prediction, highlighting their numerous uses and significant impact on healthcare delivery. Predictive models can be constructed using broad datasets and powerful machine learning techniques to anticipate the onset, development, and treatment results of many diseases, ranging from chronic ailments to infectious illnesses. The inclusion of streamlit creates an easy-to-use user experience for medical personnel and patients alike, allowing for smooth engagement with models of prediction and improving decision-making processes. In addition, the use of such prediction models has the potential to improve early detection, personalized therapy techniques, and resource management in healthcare systems. However, the application of ML-based disease prediction systems raises concerns about data privacy, model clarity, and ethical consequences. By overcoming these obstacles and seizing the opportunities provided by ML and streamlit, the future of healthcare holds enormous promise for improving patient outcomes and expanding medical research.

Print-Rich Speech Sound Therapy Sessions: The Theory Behind It and Plans for Implementation

Purpose: Print knowledge is a powerful predictor of later reading abilities, which are crucial to children's academic success. Children with speech sound disorders (SSDs) are at risk for literacy difficulties. Speech sound therapy is an opportune time to address not only speech sound production but also the connections between speech sounds and their associated letters and letter patterns. This tutorial aims to provide a rich source of evidence to support the use of print and print-referencing in speech sound therapy sessions. We include feasible suggestions for speech-language pathologists (SLPs) to facilitate print-rich experiences for children with SSDs. Incorporating orthography into therapy sessions, through activities such as shared book reading, allows SLPs to address speech sound errors while supporting children's literacy development. Conclusions: This tutorial provides an overview of the risk of literacy deficits in children with SSDs and describes the theoretical underpinnings of why incorporating print is a powerful learning tool. We expand on the research behind print referencing in preschool classrooms. Finally, we provide detailed examples, including book lists and therapy ideas for a variety of ages, to help SLPs add print referencing to their therapy toolkit.

NOVEL METHOD VITAMIN DEFICIENCY DETECTION USING ALEXNET DNN ALGORITHM

Vitamin deficiencies can have significant impacts on overall health and well-being. Early detection plays a crucial role in preventing complications and improving outcomes. However, traditional methods for detecting deficiencies can be time-consuming and costly. This project aims to develop a novel method for detecting vitamin deficiencies using the AlexNet DNN algorithm, a powerful deep learning model for image classification. The purpose of this project is to explore the feasibility of using image analysis and deep learning techniques to detect vitamin deficiencies accurately and efficiently. The objectives include improving the accuracy of detection, reducing false positives and negatives, and developing a reliable and accessible tool for early detection. To achieve our objectives, we will gather a large dataset of images depicting various vitamin deficiencies. These images will be preprocessed to enhance features and reduce noise. The AlexNet DNN algorithm will be trained on this dataset, learning to recognize patterns and features associated with different deficiencies. The algorithm will undergo rigorous testing and evaluation to ensure its effectiveness. KEYWORDS— Vitamins, Deficiency, AlexNet, Deep Neural Network(DNN), Effectiveness

Machine Learning Enabled Inventory Prediction

Effective inventory management is a critical aspect of business operations, ensuring optimal stock levels, minimizing costs, and meeting customer demand. Accurate inventory prediction plays a pivotal role in achieving these objectives. This project explores the application of the XGBoost algorithm, a powerful machine learning technique, for inventory prediction. XGBoost's ability to handle complex nonlinear relationships and its robust performance make it a promising approach for this task. The project aims to develop an inventory prediction system using the XGBoost algorithm, leveraging historical sales data and other relevant factors to forecast future inventory levels. This project demonstrates the potential of machine learning, specifically the XGBoost algorithm, to revolutionize inventory management practices, enabling businesses to achieve greater efficiency and profitability. Keyword : Machine Learning, Non-linear Relations, Cost Minimization, Stock levels, Profitability, Inventory Prediction.

Visual Intelligence: Machine Learning Approaches to Image Filtering and Identification

We apply powerful machine learning approaches in our work on visual intelligence to increase visual clarity and automate picture filtering. Machine learning is applied in image filtering in an attempt to automate and enhance picture quality, minimize noise, and extract vital information. Machine learning employs models and algorithms created on massive datasets to enhance filtering efficiency over prior techniques. Our strategy seeks to minimize noise, enhance picture quality, and extract useful information by integrating the Retinex algorithm with Gaussian filtering. Its purpose is to promote productivity and effectiveness in numerous sectors, including satellite analysis, medical imaging, and photography, by means of techniques like noise reduction, sharpening, and creative upgrades. This technology improves the area of visual data processing while solving concerns with computation efficiency and parameter fine-tuning. Furthermore, by tackling concerns like computer efficiency, dataset bias, and fine-tuning, it achieves substantial breakthroughs in the processing and interpretation of visual data. Our results hint to the potential uses of machine learning in numerous domains, such as satellite analysis, medical imaging, and photography, and illustrate how it may be used to enhance picture quality. We explain how visual intelligence is applied in current photo processing and analysis via our work. Keywords— Visual intelligence, image processing, blur removal, object detection, machine learning, Computer vision, Image enhancement, Retinex algorithm, Gaussian filtering, Deep learning, Neural networks, Feature extraction, Image segmentation, Anomaly detection, Data preprocessing.

From integers to fractions: The role of analogy in transfer and long-term learning

Fractions are the gatekeepers to advanced mathematics but are difficult to learn. One powerful learning mechanism is analogy, which builds fraction understanding on a pre-existing foundation of integer knowledge. Indeed, a short intervention that aligned fractions and integers on number lines improved children’s estimates of fractions (Yu et al., 2022). The breadth and durability of such gains, however, are unknown, and analogies to other sources (such as percentages) may be equally powerful. To investigate this issue, we randomly assigned 109 fourth and fifth graders to one of three experimental conditions with different analogical sources (integers, percentages, or fractions) or a control condition. During training, children in the experimental conditions solved pairs of aligned fraction number line problems and proportionally-equivalent problems expressed in integers, percentages, or fractions (e.g., 3/8 on a 0–1 number line aligned with 3 on a 0–8 number line). Children in the control group solved fraction number-line problems sequentially. At pretest and a two-week delayed posttest, children completed a broad fraction knowledge battery, including estimation, comparison, categorization, ordering, and arithmetic. Results showed that aligning integers and fractions on number lines facilitated better estimation of fractional magnitudes, and the training effect transferred to novel fraction problems after two weeks. Similar gains were not observed for analogies using percentages. These findings highlight the importance of building new mathematical knowledge through analogies to familiar, similar sources.