Optimal Learning Research Articles

Blockchain with optimal deep learning assisted secure data sharing and classification on future healthcare systems

In the future of healthcare, Blockchain (BC) technology holds immense potential for improving the security and privacy of data. By allowing the secure and immutable storage of medical files and healthcare-related transactions, BC ensured that sensitive medical data remains tamper-proof and open only to authorized parties. Patients have greater control over their data's development, revoking or granting access as required, but healthcare workers can streamline data sharing and ensure the integrity of important data. The decentralized nature of BC networks decreases the risk of centralized data breaches, eventually fostering trust and transparency in the healthcare ecosystems. Conversely, deep learning (DL) has great to revolutionize healthcare diagnostics in the future, offering quick and extremely accurate estimates of medical conditions. This technology has greatly enhanced patient solutions, decreased medical expenses, and improved the burden on medical staff by providing appreciated insights into an extensive range of conditions, from cancer to neurological disorders. With this stimulus, this study presents a novel BC with optimal DL-based secure data sharing and classification (BCODL-SDSC) technique in the future healthcare system. The goal of the BCODL-SDSC technique is to secure and thoroughly examine healthcare data using BC and DL techniques. Primarily, the BCODL-SDSC technique enables BC technology to store and maintain the patient’s data from the procedure of several transactions and enable access control to the various stakeholders. For the security of the medical images, the BCODL-SDSC technique applies the Fractional Order Lorenz system (FOLS) based encryption technique with tuna swarm optimization (TSO) algorithm based optimal key generation process. Finally, a multi-stage process performs the classification of the medical images: MobileNetv1 feature extractor, artificial rabbit’s optimization (ARO) based hyperparameter tuning, and stacked recurrent neural network (SRNN) based classification. The experimental outcome of the BCODL-SDSC technique was examined on a benchmark medical image database. An extensive comparative study reported that the BCODL-SDSC technique reaches an effective performance with other models with a maximum accuracy of 99.11%.

LSTM network optimization and task network construction based on heuristic algorithm

This work aims to advance the security management of complex networks to better align with evolving societal needs. The work employs the Ant Colony Optimization algorithm in conjunction with Long Short-Term Memory neural networks to reconstruct and optimize task networks derived from time series data. Additionally, a trend-based noise smoothing scheme is introduced to mitigate data noise effectively. The approach entails a thorough analysis of historical data, followed by applying trend-based noise smoothing, rendering the processed data more scientifically robust. Subsequently, the network reconstruction problem for time series data originating from one-dimensional dynamic equations is addressed using an algorithm based on the principles of Stochastic Gradient Descent (SGD). This algorithm decomposes time series data into smaller samples and yields optimal learning outcomes in conjunction with an adaptive learning rate SGD approach. Experimental results corroborate the remarkable fidelity of the weight matrix reconstructed by this algorithm to the true weight matrix. Moreover, the algorithm exhibits efficient convergence with increasing data volume, manifesting shorter time requirements per iteration while ensuring the attainment of optimal solutions. When the sample size remains constant, the algorithm’s execution time is directly proportional to the square of the number of nodes. Conversely, as the sample size scales, the SGD algorithm capitalizes on the availability of more information, resulting in improved learning outcomes. Notably, when the noise standard deviation is 0.01, models predicated on SGD and the Least-Squares Method (LSM) demonstrate reduced errors compared to instances with a noise standard deviation of 0.1, highlighting the sensitivity of LSM to noise. The proposed methodology offers valuable insights for advancing research in complex network studies.

Looking back to move forward: comparison of instructors’ and undergraduates’ retrospection on the effectiveness of online learning using the nine-outcome influencing factors

This study delves into the retrospections of undergraduate students concerning their online learning experiences after the COVID-19 pandemic, using the nine key influencing factors: behavioral intention, instruction, engagement, interaction, motivation, self-efficacy, performance, satisfaction, and self-regulation. 46 Year 1 students from a comprehensive university in China were asked to maintain reflective diaries throughout an academic semester, providing first-person perspectives on the strengths and weaknesses of online learning. Meanwhile, 18 college teachers were interviewed with the same questions as the students. Using thematic analysis, the research identified 9 factors. The research revealed that instruction ranked highest among the 9 factors, followed by engagement, self-regulation, interaction, motivation, and others. Moreover, teachers and students had different attitudes toward instruction. Thirdly, teacher participants were different from student participants given self-efficacy and self-regulation due to their variant roles in online instruction. Lastly, the study reflected students were not independent learners, which explained why instruction ranked highest in their point of view. Findings offer valuable insights for educators, administrators, and policy-makers involved in higher education. Recommendations for future research include incorporating a more diverse sample, exploring relationships between the nine factors, and focusing on equipping students with skills for optimal online learning experiences.

Prediction of Glass Chemical Composition and Type Identification Based on Machine Learning Algorithms

Ancient glass artifacts were susceptible to weathering from the environment, causing changes in their chemical composition, which pose significant obstacles to the identification of glass products. Analyzing the chemical composition of ancient glass has been beneficial for evaluating their weathering status and proposing measures to reduce glass weathering. The objective of this study was to explore the optimal machine learning algorithm for glass type classification based on chemical composition. A set of glass artifact data including color, emblazonry, weathering, and chemical composition was employed and various methods including logistic regression and machine learning techniques were used. The results indicated that a significant correlation (p < 0.05) could only observed between surface weathering and the glass types (high-potassium and lead–barium). Based on the random forest and logistic regression models, the primary chemical components that signify glass types and weathering status were determined using PbO, K2O, BaO, SiO2, Al2O3, and P2O5. The random forest model presented a superior ability to identify glass types and weathering status, with a global accuracy of 96.3%. This study demonstrates the great potential of machine learning for glass chemical component estimation and glass type and weathering status identification, providing technical guidance for the appraisal of ancient glass artifacts.

Applications of Alpha Neurofeedback Processes for Enhanced Mental Manipulation of Unfamiliar Molecular and Spatial Structures

Applications of Alpha Neurofeedback Processes for Enhanced Mental Manipulation of Unfamiliar Molecular and Spatial Structures

An Intelligent Framework for English Teaching through Deep Learning and Reinforcement Learning with Interactive Mobile Technology

As globalization deepens, the significance of English teaching in the educational landscape has become more prominent. Traditional teaching methods are increasingly inadequate for providing personalized and efficient learning experiences. This gap is being addressed by the rapid advancements in artificial intelligence, especially through deep and reinforcement learning. These technologies provide a framework for intelligent English teaching systems by mimicking human learning processes to customize personalized learning experiences, optimize learning paths, and enhance efficiency. However, challenges remain in fine-tuning teaching strategies to meet the varying needs of individual learners and dynamically adapting to their evolving interests in the short term. This study introduces a novel framework for an intelligent English teaching system that leverages the potential of interactive mobile technology alongside a deep Q-network (DQN) algorithm to dynamically adjust English teaching strategies. This approach enables real-time personalization of teaching strategies to create optimal learning paths for individual learners. Moreover, it incorporates a model based on neural collaborative filtering to capture and adapt to learners’ short-term dynamic interests, thereby recommending relevant learning content in real-time. This framework enhances learning efficiency and personalizes content delivery, demonstrating considerable theoretical and practical value for the future of educational technology.

Designing a smart classroom with advanced Audiovisual equipment and soft collaboration capabilities for Higher Education

This is a technology-enabled classroom design which reflects a thoughtful integration of cutting-edge audiovisual solutions to enhance the learning environment. The inclusion of automated audio-visual solutions indicates a commitment to creating a seamless and efficient learning experience through technology. The incorporation of a voice lift application ensures clear and audible communication for the instructor, addressing potential challenges in large classroom settings. The use of Dante-based communication over Ethernet signifies a high-quality and reliable audio distribution system, enabling efficient communication within the classroom. The inclusion of features catering to individuals with disabilities demonstrates a commitment to creating an inclusive and accessible learning environment. The beamforming microphone array system with multiple elements is a sophisticated technology that allows for effective tracking of the active participant. This ensures that the audio focus adapts to the speaker's location. The automatic and steerable coverage type enhances the adaptability of the microphone array system, automatically focusing on the speaker. This feature contributes to an immersive and dynamic learning experience. The support for Power over Ethernet (PoE) simplifies the installation and management of AV devices in the classroom. It reduces the need for additional power outlets and cables, streamlining the overall setup. Overall, this design prioritizes advanced audiovisual technologies to create a unique, state-of-the-art, and inclusive technology-enabled classroom. The focus on clear communication, adaptability, and ease of use demonstrates a commitment to providing an optimal learning environment for both faculty and students

A new method of rapid-intelligent assessment for the dynamic service performance in high-speed railway subgrade

ABSTRACT There is a lack of a rapid, advanced, and comprehensive method for assessing the dynamic service performance in high-speed railway subgrade, which leads to an inability to pre-assess the health state of the subgrades in advance. In this paper, an innovative method of rapid-intelligent assessment for the dynamic service performance (RADSP) in high-speed railway subgrade was proposed. Firstly, a large-scale exciter was used to simulate the train loading effect, which could rapidly obtain the assessment metrics for subgrade dynamic service performance, including subgrade dynamic response (SDR) and subgrade cumulative deformation (SCD), where the SDR includes dynamic stress and acceleration. Then, to explore the future development trend of SCD during operation, an intelligent and advanced prediction of SCD was conducted using the optimal machine learning (ML) prediction model. Finally, by combining the results of the SDR and SCD assessments to achieve the RADSP in high-speed railway subgrade. Moreover, the method was applied to assess the dynamic service performance of diatomaceous earth subgrade structures with four different waterproofing structure layers (WSL): the middle of subgrade surface layer (WSL-1), the bottom of subgrade surface layer (WSL-2), the middle of subgrade bottom layer (WSL-3), and the bottom of subgrade bottom layer (WSL-4). The results indicate that: 1) During the excitation experiments, the maximum values for SDR (dynamic stress, acceleration) and SCD exceed the respective limits of 50 kPa, 10 m/s2, and 15 mm under WSL-1 and WSL-4. 2) The ensemble model, AdaBoost-GPR, is selected as the optimal ML model to predict the SCD values. Notably, the prediction SCD value exceeds the 15 mm limit under WSL-3, indicating that the SCD during future operation does not meet the requirements for pre-assessment. 3) Under WSL-2, both SDR and SCD assessments satisfy the limit value requirements, indicating that the dynamic service performance is better. Based on the proposed method, it is found that when designing waterproof-drainage subgrade structures in diatomaceous earth areas, it is recommended to position the WSL at the bottom of subgrade surface layer. The research findings serve as a crucial reference and valuable lesson for the health monitoring of high-speed railway subgrades.

Bioinformatics analysis of PANoptosis regulators in the diagnosis and subtyping of steroid-induced osteonecrosis of the femoral head.

In this study, we aimed to investigate the involvement of PANoptosis, a form of regulated cell death, in the development of steroid-induced osteonecrosis of the femoral head (SONFH). The underlying pathogenesis of PANoptosis in SONFH remains unclear. To address this, we employed bioinformatics approaches to analyze the key genes associated with PANoptosis. Our analysis was based on the GSE123568 dataset, allowing us to investigate both the expression profiles of PANoptosis-related genes (PRGs) and the immune profiles in SONFHallowing us to investigate the expression profiles of PRGs as well as the immune profiles in SONFH. We conducted cluster classification based on PRGs and assessed immune cell infiltration. Additionally, we used the weighted gene co-expression network analysis (WGCNA) algorithm to identify cluster-specific hub genes. Furthermore, we developed an optimal machine learning model to identify the key predictive genes responsible for SONFH progression. We also constructed a nomogram model with high predictive accuracy for assessing risk factors in SONFH patients, and validated the model using external data (area under the curve; AUC = 1.000). Furthermore, we identified potential drug targets for SONFH through the Coremine medical database. Using the optimal machine learning model, we found that 2 PRGs, CASP1 and MLKL, were significantly correlated with the key predictive genes and exhibited higher expression levels in SONFH. Our analysis revealed the existence of 2 distinct PANoptosis molecular subtypes (C1 and C2) within SONFH. Importantly, we observed significant variations in the distribution of immune cells across these subtypes, with C2 displaying higher levels of immune cell infiltration. Gene set variation analysis indicated that C2 was closely associated with multiple immune responses. In conclusion, our study sheds light on the intricate relationship between PANoptosis and SONFH. We successfully developed a risk predictive model for SONFH patients and different SONFH subtypes. These findings enhance our understanding of the pathogenesis of SONFH and offer potential insights into therapeutic strategies.

Security Development Lifecycle-based Adaptive Reward Mechanism for Reinforcement Learning in Continuous Integration Testing Optimization

Security Development Lifecycle-based Adaptive Reward Mechanism for Reinforcement Learning in Continuous Integration Testing Optimization

Risk-averse Reinforcement Learning for Portfolio Optimization

This investigation explores Reinforcement Learning (RL) for dynamic portfolio optimization with risk assessment. The challenges include market complexity, uncertain reactions, and regulatory requirements for risk-averse decisions. Our solution leverages Bayesian Neural Network (BNN) to capture uncertainties. We successfully implemented a risk-averse Reinforcement Learning algorithm, achieving 18 percent lower risk. Reinforcement Learning with risk-aversion shows promise for optimizing portfolios for risk-averse investors.

Towards a Digital Twin Framework in Additive Manufacturing: Machine Learning and Bayesian Optimization for Time Series Process Optimization

Laser directed-energy deposition (DED) offers notable advantages in additive manufacturing (AM) for producing intricate geometries and facilitating material functional grading. However, inherent challenges such as material property inconsistencies and part variability persist, predominantly due to its layer-wise fabrication approach. Critical to these challenges is heat accumulation during DED, influencing the resultant material microstructure and properties. Although closed-loop control methods for managing heat accumulation and temperature regulation are prevalent in DED literature, few approaches integrate real-time monitoring, physics-based modeling, and control simultaneously in a cohesive framework. To address this, we present a digital twin (DT) framework for real-time model predictive control of process parameters of the DED for achieving a specific process design objective. To enable its implementation, we detail the development of a surrogate model utilizing Long Short-Term Memory (LSTM)-based machine learning which uses Bayesian Inference to predict temperatures across various spatial locations of the DED-built part. This model offers real-time predictions of future temperature states. In addition, we introduce a Bayesian Optimization (BO) method for Time Series Process Optimization (BOTSPO). Its foundational principles align with traditional BO, and its novelty lies in our unique time series process profile generator with a reduced dimensional representation. BOTSPO is used for dynamic process optimization in which we deploy BOTSPO to determine the optimal laser power profile, aiming to achieve desired mechanical properties in a DED build. The identified profile establishes a process trajectory that online process optimizations aim to match or exceed in performance. This paper elucidates components of the digital twin framework, advocating its prospective consolidation into a comprehensive digital twin system for AM.

Integrating few-shot personalized thermal comfort model and reinforcement learning for HVAC demand response optimization

Heating, ventilation, and air conditioning (HVAC) is the most common equipment to maintain personalized thermal comfort (PTC), and a valuable demand response (DR) resource. The challenges to be addressed in HVAC DR optimization include uncertainties of environment, PTC varies by users and rapid decision-making. This paper proposed a fast HVAC DR optimization method that integrates deep reinforcement learning (DRL) and few-shot PTC model. Firstly, we established an HVAC DR optimization framework integrated few-shot PTC model and DRL, in which the few-shot PTC model was utilized to generate user’s PTC temperature range served as the constraint and state variable of DRL. Secondly, we proposed a few-shot PTC model, which only required several field samples and had excellent engineering usability. Thirdly, we designed the HVAC DR optimization algorithm considering user’s PTC based on proximal policy optimization, which automatically dealt with the uncertainties and made rapid decisions for setting HVAC setpoints of temperature dynamically. Simulation results show that the few-shot PTC model only requires 9 field samples, which is less than 1% of existing advanced methods. Furthermore, the proposed method converges within 12.6 min under uncertainties on an i7-12700 processor, which is only about 1/2 of the standard deep deterministic policy gradient.

An experimental investigation and predictive modeling using machine learning technique for reclamation of metal values from scrap NdFeB magnets

Neodymium-Iron-Boron (NdFeB) magnets are widely used in various industries due to their exceptional magnetic properties, such as high coercivity, remanence, and maximum energy product. These magnets consist of rare earth elements (REEs) viz., neodymium (Nd), praseodymium (Pr), and Dysprosium (Dy), along with other metals. With the ever increasing demand for REEs and the need to bridge the supply gap, it is crucial to develop alternative methods for their extraction. Recycling metal from scrap magnets is a promising approach to address the pressure on the supply chain and work towards sustainable development. The primary goal of this study is to generate a predictive model based on machine learning to determine the optimal conditions for metal recovery from scrap NdFeB magnets through water leaching after chloridizing roasting. Bench-scale leaching experiments were carried out to generate a dataset for statistical process optimization and machine learning analysis. The leaching kinetics of neodymium was also explored, and mixed-controlled shrinking core model was found to be most suitable, with an activation energy of 58.11 kJ/mol in the temperature range of 25– 95 °C. This study is the first to utilize a machine learning approach to analyze the potential process variables and their impact on metal recovery from calcined NdFeB magnet powder. A comparative analysis between experimental and machine learning approaches is presented to predict the optimal conditions for selective recovery of metal values from scrap NdFeB magnets. The maximum efficiency of extraction of metal ions was found to occur at a temperature of 95 °C, solid to liquid ratio of 125 g/l, and leaching duration of 60 min.

The prognostic role of an optimal machine learning model based on clinical available indicators in HCC patients

The prognostic role of an optimal machine learning model based on clinical available indicators in HCC patients

Machine-learning-aided regional post-seismic usability prediction of buildings: 2016-2017 Central Italy earthquakes

This study addresses the development of machine learning models for predicting the post-seismic buildings usability within regions prone to frequent earthquakes. The analysis leverages on field data from 2016-2017 Central Italy earthquakes. Several machine learning techniques are employed for this task, namely K-nearest neighbors, linear support vector machine, radial basis function support vector machine, decision tree, random forest, neural network, adaptive boosting, naive Bayes, quadratic discriminant analysis, logistic regression, and linear discriminant analysis. The input variables include both building attributes and seismic intensity measures. Since the database turns out to be strongly imbalanced, the potential influence of two preprocessing techniques is examined, namely principal component analysis and synthetic minority oversampling technique. Several metrics are considered to evaluate the performance of the resulting predictive machine learning models. Moreover, this study investigates the optimal machine learning model’s robustness against uncertainties, quantifies the importance of its features, and investigates how usability classes clustering can impact its performance. Every step of the implemented procedure is deeply explained and discussed to provide useful guidelines for similar applications.

Portfolio Optimization with Prediction-Based Return Using Long Short-Term Memory Neural Networks: Testing on Upward and Downward European Markets

AbstractIn recent years, artificial intelligence has helped to improve processes and performance in many different areas: in the field of portfolio optimization, the inputs play a crucial role, and the use of machine learning algorithms can improve the estimation of the inputs to create robust portfolios able to generate returns consistently. This paper combines classical mean–variance optimization and machine learning techniques, concretely long short-term memory neural networks to provide more accurate predicted returns and generate profitable portfolios for 10 holding periods that present different financial contexts. The proposed algorithm is trained and tested with historical EURO STOXX 50® Index data from January 2015 to December 2020, and from January 2021 to June 2022, respectively. Empirical results show that our LSTM neural networks are able to achieve minor predictive errors since the average of the MSE of the 10 holding periods is 0.00047, the average of the MAE is 0.01634, and predict the direction of returns with an average accuracy over the 10 investment periods of 95.8%. Our prediction-based portfolios consistently beat the EURO STOXX 50® Index, achieving superior positive results even during bear markets.

The influence of flipbook learning media, learning interest, and learning motivation on learning outcomes

The aim of this research is to investigate the influence of flipbook learning media, learning interest, and learning motivation on junior high school students' learning outcomes. The method used is the regression method with a quantitative approach. This research was conducted at Junior High School 1 Yogyakarta with a sample of 64 class VIII social studies students. Data collection consists of interviews, observations, and documentation. Prerequisite test analysis consists of tests for normality, multicollinearity, and heteroscedasticity. Hypothesis testing using simple regression, and multiple regression. The research results show that there is an influence: i) flipbook learning media on learning outcomes with a tcount of 73.33, a significance value of 0.000 0.005, ii) interest in learning on learning outcomes with a tcount of 33.678, a significance value of 0.000 0.005, iii) learning motivation on learning outcomes with a tcount of 30.678, a significance value of 0.000 0.005, and iv) flipbook learning media, learning interest, and learning motivation together on learning outcomes with Fcount 47.879 Ftable 2.77 with a significance of 0.000 0.005. The conclusion is that the use of flipbook learning media, increasing interest in learning, and strengthening learning motivation can support each other to achieve optimal learning outcomes for students.

Infrared small target detection algorithm based on filter kernel combination optimization learning method

Infrared small target detection in complex backgrounds remains a big challenge due to its small size and weak energy, which makes it easy to be submerged in complex backgrounds. Common filtering algorithms are difficult to apply to various complex scenes and generate many false alarms and poor robustness. Therefore, this paper proposes an infrared small target detection algorithm based on filter kernel combination optimization learning method. For the sample image block use typical filter kernels and weight coefficients to calculate the combination of filtering results, and construct the loss function of the results and the true value. The optimal weight coefficients are learned using an optimization method, which effectively achieves the fusion of multiple features, then we build a sample library containing comprehensive priori information. In the implementation, an online inference method based on similarity measure is proposed to match test images with sample library image blocks to obtain optimal combination weights. By utilizing the comprehensive priori information to enhance the adaptive ability of the algorithm for various complex backgrounds. Finally, get the position of the infrared small target. The experimental results show that our algorithm has excellent detection performance for infrared small targets with complex background and low signal to clutter ratio, and the AUC value is 0.9980, which is more robust than other comparative algorithms.

Mean squared error bound for learning-based multi-target localization and its application in learning network architecture design

The learning-based method has been widely applied in target localization. However, the performance of learning-based localization is mostly validated by simulation experiments and lacks theoretical analysis. In this paper, we investigate the performance of learning-based multi-target localization in a multiple-input multiple-output (MIMO) radar system with widely separated antennas. A mean squared error bound (MSEB) for learning-based multi-target localization is derived, which contains all the system parameters, including radar-related parameters and learning network-related parameters. Therefore, the MSEB can be employed as the metric to optimize any parameter of interest, such as those associated with the learning network architecture, which is usually determined with the trial-and-error method by minimizing the training loss. Leveraging the MSEB, we can design the optimal learning network architecture which maximizes the testing performance and this optimal architecture can be obtained before training. Numerical results are presented to verify the correctness of the MSEB and effectiveness of the MSEB-based learning network architecture design method.