Search

This paper presents findings on machine learning agent behavior prediction in a board game application developed by a group of students. The goal of this research is to create a model facilitating collaboration between a user and an AI to play together in the board game using a Human-in-the-Loop architecture. By injecting explainability, the aim is to enhance communication and understanding between the user and the AI agent. Featuring a competitive Artificial Intelligence (AI) based on the Proximal Policy Optimization model, this research explores methods to make AI decisions transparent for enhanced player understanding. Two predictive models, a Decision Tree (DT) and a Deep Learning (DL) classifier, were developed and compared. The results show that the DT model is effective for short-term predictions but limited in broader applications, while the DL classifier shows potential for long-term prediction without requiring direct access to the game's AI. This study contributes to understanding human-AI interaction in gaming and offers insights into AI decision-making processes. [ABSTRACT FROM AUTHOR]

IoT technologies are becoming more and more common in our daily activities because the networks they create are capable of collecting information, monitoring and controlling remotely. However, these devices are not exempt from security attacks, as they become vulnerable entry points to data networks. The use of traditional methods to secure networks (e.g., Next Generation Firewalls (NGFW), encryption, etc.) is not recommended because the devices used in this type of network are limited in terms of computing power and storage availability (e.g., nodeMCU). In this paper, we propose to design two intrusion detection systems in embedded systems using machine learning (ML) algorithms, Artificial Neural Networks and K-means. In a distributed architecture Wireless Sensor Network scenario (WSN), we evaluate their performance in terms of connection and response times, detection accuracy and intruder detection time. Simulation results show that both models are able to find irregularities in network traffic within milliseconds. [ABSTRACT FROM AUTHOR]

Every data packet must pass through a few intermediate nodes to reach its destination. Among other reasons, tremendous growth in internet devices encourages those intermediate nodes to drop the data packets. Optimizing the data packet route is an effective solution to deal with packet loss. Advanced machine learning approaches have been identified as a powerful support tool for routing optimization in node networks. Cloud computing has kept pace with the continuously developing hardware infrastructure. Improved connection, processing power, and memory units enable real-time machine learning. This paper proposes and evaluates an approach for optimizing the packet path, by one hop, for intermediate nodes as a backup called Cloud Acknowledgement Scheme (CAS). It offers information on the transmission trend and the tendencies of certain adjacent nodes or groups of neighboring nodes in a network. The proposed CAS has been validated via a series of machine learning experiments using real-world node data. [ABSTRACT FROM AUTHOR]

This paper presents a novel approach for measuring multi-shapes degrees of similarities. A new shape transformation concept is suggested through mapping the closed boundary of the shape using unfolded process into one-to-one equivalent graph (or signal) where ample number of approaches for similarity testing can be applied. The degree of similarity is then measured by calculating the discrepancies between each of the two transformed unfolded-graphs autocorrelation functions. The proposed method is also effective in applying clustering technique for clustering groups of shapes into three hierarchical similarity levels depending on their mutual degree of similarity. The suggested similarity method handles two categories of shapes: geometric and non-geometric. Unlike other well-known techniques in the subject of machine learning and deep learning, the presented method represents a powerful analytical-based technique for similarity analysis of both regular and irregular shapes regardless of size or orientation, and in effectively handling compound or multi-shapes degrees of similarities. The results of testing the approach on selected datasets have successfully demonstrated the great effectiveness of the developed technique in yielding output multi-shape degrees of similarities in new graphical and comprehensive forms. Finally, it is pointed out that the new approach will have many potential real-life applications such as in industry, medicine, biology, security, and authentications. [ABSTRACT FROM AUTHOR]

A neural network is a machine learning (ML) program or model that processes information and recognizes patterns, similar to the human brain. The neural network algorithm operates by training on data to learn and enhance its accuracy. If there is any mistake in the learning process, the machine will react unexpectedly and produce incorrect information. So, whenever a neural network model is developed, it is mandatory to evaluate its performance and ensure its output is accurate. Cameras, touchscreens, internet connectivity, and powerful CPUs have contributed to the popularity of smartphones. Mobile apps are software applications that run on smartphones. ML enhances mobile app functionality by offering features such as voice recognition, image analysis, natural language processing, personalization, and recommendations. Training ML models using mobile apps is challenging due to limited resources, data, and privacy concerns. Object recognition is a neural network-based technique that enables the identification and localization of objects in an image or video. This technique is used in driverless cars, disease identification, industrial inspection, robotics, and more. In this paper, the author introduces a new neural network-based algorithm that utilizes mobile devices for training and recognizing images. Although mobile devices have some technological limitations for training, well-established guidelines for systematically mapping verification and validation techniques have been used in the proposed neural network to ensure performance and correctness in on-device object training and recognition. [ABSTRACT FROM AUTHOR]

Due to its flexibility, accessibility and the increasing importance of digital literacy, online learning has gained priority in the last few years. However, several challenges have led students to resist it. Predicting students’ attitudes towards online learning could assist educators and educational institutions in addressing these challenges and enhancing its effectiveness. This paper presents a Learning Attitude Prediction Model (LAPM) that can detect students’ attitudes from their informal Arabic texts. To generate the desired LAPM, five machine learning algorithms and three different approaches for text representation are employed. In addition, handling stop words is another important issue when dealing with informal Arabic text. Two scenarios are commonly adopted: preserving and eliminating stop words. The best result was obtained when using a support vector machine (SVM) classifier coupled with the term frequency-inverse document frequency (TF-IDF) approach and preserving stop-words, achieving an F1-score of 85.4%. Therefore, an effective LAPM could be developed to predict students’ attitudes. Using LAPM, educators and educational institutions can monitor students’ attitudes toward online learning and provide personalized support to individual students. Consequently, an enhancement in student satisfaction and an improvement in academic achievement could be achieved. [ABSTRACT FROM AUTHOR]

Forecast accuracy is a crucial topic for industrial companies, and its impacts are particularly important for the finance and production departments. The company can incur high costs if forecasts are inaccurate, for example, due to stock-outs or excess inventory. Therefore, this study aimed to optimize accessories forecasting for a medium-sized Swiss enterprise. To do so, different forecasting techniques were tested, and statistical methods and machine learning (ML) algorithms were compared. The results were adjusted according to key account managers' (KAM) expertise. This paper presents a comparison between exponential smoothing, seasonal autoregressive integrated moving average (SARIMA), SARIMAX (SARIMA with exogenous variables) and ML algorithms, such as knearest neighbors (k-NN), least absolute shrinkage and selection operator (LASSO) regression, linear regression, and even random forest (RF). To compare these different methods, two measures of statistical dispersion are computed: mean absolute error (MAE) and root mean squared error (RMSE). The results are standardized to enable a better comparison. For our dataset, SARIMAX (with the KAMs' expertise as an exogenous variable) gives better results than all the ML algorithms tested. [ABSTRACT FROM AUTHOR]

Functional End-of-Line Testing is a powerful but expensive approach for industrial quality assurance. A major cost driver of End-of-Line Testing is the often overlong and overcomplex design of industrial test cycles. To tackle this challenge, the paper proposes a data-driven approach based on machine learning for the optimization of the End-of-Line test cycle of a hydraulic pump. As each test cycle comprises a multivariate and multi-phased time series, the optimization can be interpreted as a form of feature selection that aims to select the minimal subset of relevant and non-redundant testing sensors and cycle phases. We define the problem as a multi-objective optimization task balancing the goals of maximizing the test cycle sensitivity and specificity while minimizing the test cycle complexity. The general idea of the optimization approach is inspired by the iterative concept of wrapper feature selection. Our approach was validated in an experiment by applying it to real-world End-of-Line Testing data of an hydraulic axial piston pump. The results have shown that the approach could be successfully used to find an improved test cycle design for the pump and, especially, Backward Elimination performed well as solution search algorithm. [ABSTRACT FROM AUTHOR]

In the traditional well log depth matching tasks, manual adjustments are required, which means significantly labor-intensive for multiple wells, leading to low work efficiency. This paper introduces a multi-agent deep reinforcement learning (MARL) method to automate the depth matching of multi-well logs. This method defines multiple top-down dual sliding windows based on the convolutional neural network (CNN) to extract and capture similar feature sequences on well logs, and it establishes an interaction mechanism between agents and the environment to control the depth matching process. Specifically, the agent selects an action to translate or scale the feature sequence based on the double deep Q-network (DDQN). Through the feedback of the reward signal, it evaluates the effectiveness of each action, aiming to obtain the optimal strategy and improve the accuracy of the matching task. Our experiments show that MARL can automatically perform depth matches for well-logs in multiple wells, and reduce manual intervention. In the application to the oil field, a comparative analysis of dynamic time warping (DTW), deep Q-learning network (DQN), and DDQN methods revealed that the DDQN algorithm, with its dual-network evaluation mechanism, significantly improves performance by identifying and aligning more details in the well log feature sequences, thus achieving higher depth matching accuracy. [ABSTRACT FROM AUTHOR]

In the paper, an intelligent actor-critic learning control method augmented by a fuzzy broad learning system with output recurrent feedback (abbreviated as ORFBLS) is proposed for obstacle-avoiding trajectory tracking of heterogeneous Mecanum-wheeled omnidirectional mobile robots (HMOMRs) augmented with un-modeled errors and the learning control system's parameters are changed with the working environment. By using actor-critic learning algorithms and the ORFBLS online function approximation, a new ORFBLS-based actor-critic learning control approach, dubbed as ORFBLS-ACLC, is presented to accomplish stable trajectory tracking of such uncertain HMOMRs. The ORFBLS-ACLC method combined with an obstacle avoidance scheme is presented to follow the desired collision-free trajectory tracking more closely than the existing methods do. The merits, superiority, and practicality of the raised control approach are well verified by carrying out three simulation comparisons and two experiments on an uncertain HMOMR. [ABSTRACT FROM AUTHOR]

In view of the problems existing in the path planning algorithms of unmanned vehicles, such as low search efficiency, slow convergence speed and easy to fall into the local optimal. Based on the characteristics of route planning for unmanned vehicles, this paper introduces Q-Learning into the traditional ant colony algorithm to enhance the learning ability of the algorithm in dynamic environment, so as to improve the overall efficiency of route search. By mapping pheromones into Q values in Q-learning, rapid search in complex environments is realized, and a collection-free path satisfying constraints is quickly found. The results of case analysis show that compared with the traditional ant colony algorithm and the improved ant colony algorithm considering reward and punishment factors, the improved ant colony algorithm based on Q-Learning can effectively reduce the number of iterations, shorten the path optimization time and path length and other performance indicators, and has many advantages in jumping out of the local optimal, improving the global search ability and improving the convergence speed, and has good adaptability and robustness in complex environments. It ensures the safety and stability of unmanned vehicles in complex environments. [ABSTRACT FROM AUTHOR]

Copyright of Informatica (03505596) is the property of Slovene Society Informatika and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

Copyright of Informatica (03505596) is the property of Slovene Society Informatika and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

Oracle bones inscription is an important part of Chinese cultural heritage, and the automatic segmentation technology of oracle bones is of great significance to promote the research of oracle bones. The main challenges for automatic segmentation of oracle bone inscriptions include the interference of complex backgrounds, the extraction of text regions of tiny sizes, and the problems of noise and cracks in images, which are hard to deal with using traditional image processing and machine learning techniques. With the development of artificial intelligence technology, especially the breakthrough of deep learning and convolutional neural network (CNN) in image recognition, new solutions are provided for automatic segmentation of oracle bone inscriptions. In this paper, we introduce the YOLOv8 model to the automatic segmentation task of oracle bone images. The experimental results show that the YOLOv8 model achieves satisfactory performance in this task, which verifies the effectiveness of advanced deep learning models, and may provide technical support for the digitization and cultural inheritance of oracle bones in future work. [ABSTRACT FROM AUTHOR]

ESG ratings, as a metric for assessing corporate social responsibility, have garnered escalating interest from both domestic and international investors within China. However, critical indicators for augmenting ESG scores remain underexplored. Utilizing data from 2017 to 2020 on CSI 300 and CSI 500 constituent stocks, this study employs advanced machine learning algorithms, including XGBoost (XGB), LightGBM (LGB), and Random Forest (RF), to conduct a quantitative analysis of corporate feature values and their correlation with institutional rating outcomes. The findings indicate a predominant influence of corporate governance indicators, particularly in the realm of information disclosure, followed by social responsibility and environmental metrics. The paper further delineates specific, actionable short-term and long-term strategies that corporations can adopt across environmental, social, and governance dimensions, tailored to the nuances of various sub-indicators. [ABSTRACT FROM AUTHOR]

Modern society has accumulated great amounts of knowledge since human start learning and communicating. With knowledge's continuously thriving and getting refined, knowledge itself has developed into a complex system with a huge number of content and structure features. Together with the development of science and technology, the ways of people learning have been changed greatly. Under the pressure of high-paced life and workplace competition, more and more people are in demand of personalized learning paths, in order to achieve their learning target as fast as possible. A personized and scientific learning path is the key to effective learning grogram. However, studies on algorithms of optimal learning path discovery are not comprehensively developed, especially on its performance and application. Therefore, in this article, we propose a new algorithm, Portfolio ST, which is improved from Steiner Tree, for personalized and optimal learning path mining. At first, this paper constructs the knowledge graph of finance, then builds learner models by profiling learners' owned and aimed knowledge based on the knowledge network. At last, according to the knowledge structure of a person's learner model, the most personized and shortest learning path is calculated by the Portfolio ST. After the experiment compared with two baseline algorithms, the Portfolio ST algorithm is proved to have the best performance and scalability, by changing owned/aimed/overall knowledge respectively, which is also the best algorithm of learning path mining that can achieve the global optimal among these three algorithms. [ABSTRACT FROM AUTHOR]

Graph anomaly detection (GAD) has been extensively studied in recent years. GAD aims to detect nodes, edges, and subgraphs that exhibit characteristics and distributions different from those of the majority of graph data. With the advancement of deep learning, many researchers have applied machine learning to address anomaly detection at various scales. In this paper, we classify GAD methods into detector-based and classifer-based approaches and provide a brief introduction and summary of relevant articles from the past three years. Finally, we analyze the challenges and future development directions in the field of GAD. [ABSTRACT FROM AUTHOR]

In this paper, we explore how to use a Nvidia Jetson Nano and Python to create a system that detects weariness in a person's face using computer vision and machine learning techniques. The system captures the person's face in real time, preprocesses the picture to cut noise, then detects the face using Haar cascades. Next, using computer vision algorithms, characteristics associated with weariness, such as the eyes, are retrieved. These characteristics are then used to build a machine learning model that can predict weariness in the live stream feed. Lastly, using a graphical user interface, the findings are shown, and the system may be fine-tuned to increase accuracy. This device might be used in applications such as driver monitoring and traffic safety. [ABSTRACT FROM AUTHOR]

Copyright of Journal of Donghua University (English Edition) is the property of Journal of Donghua University Editorial Board and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

With the evolving complexity of connected vehicle features, the volume and diversity of data generated during driving continue to escalate. Enabling data sharing among interconnected vehicles holds promise for improving users' driving experiences and alleviating traffic congestion. Yet, the unintentional disclosure of users' private information through data sharing poses a risk, potentially compromising the interests of vehicle users and, in certain cases, endangering driving safety. Federated learning (FL) is a newly emerged distributed machine learning paradigm, which is expected to play a prominent role for privacy-preserving learning in autonomous vehicles. While FL holds significant potential to enhance the architecture of the Internet of Vehicles (IoV), the dynamic mobility of vehicles poses a considerable challenge to integrating FL with vehicular networks. In this paper, a novel clustered FL framework is proposed which is efficient for reducing communication and protecting data privacy. By assessing the similarity among feature vectors, vehicles are categorized into distinct clusters. An optimal vehicle is elected as the cluster head, which enhances the efficiency of personalized data processing and model training while reducing communication overhead. Simultaneously, the Local Differential Privacy (LDP) mechanism is incorporated during local training to safeguard vehicle privacy. The simulation results obtained from the 20newsgroups dataset and the MNIST dataset validate the effectiveness of the proposed scheme, indicating that the proposed scheme can ensure data privacy effectively while reducing communication overhead. [ABSTRACT FROM AUTHOR]

Wireless Capsule Endoscopy (WCE) is a medical diagnostic technique recognized for its minimally invasive and painless nature for the patients. It uses remote imaging techniques to explore various segments of the gastrointestinal (GI) tract, particularly the hard-to-reach small intestine, making it an effective alternative to traditional endoscopic techniques. However, physicians face a significant challenge when it comes to analyzing a large number of endoscopic images due to the effort and time required. It is therefore imperative to implement aided-diagnostic systems capable of automatically detecting suspicious areas for subsequent medical assessment. In this paper, we present a novel approach to identify gastrointestinal tract abnormalities from WCE images, with a particular focus on ulcerated areas. Our approach involves the use of the Median Robust Extended Local Binary Pattern (MRELBP) descriptor, which effectively overcomes the challenges faced when WCE image acquisition, such as variations in illumination and contrast, rotation, and noise. Using machine learning algorithms, we conducted experiments on the extensive Kvasir-Capsule dataset, and subsequently compared our results with recent relevant studies. Noteworthy is the fact that our approach achieved an accuracy of 97.04% with the SVM (RBF) classifier and 96.77% with the RF classifier. [ABSTRACT FROM AUTHOR]

In this paper we present a research-through-design study where we employed text-to-text, text-to-image, and image-to-image generative tools for a conceptual architecture project for the eVolo skyscraper competition. We trained these algorithms on a dataset that we collected and curated, consisting of texts about and images of architecture. We describe our design process, present the final proposal, reflect on the usefulness of such tools for early-stage design, and discuss implications for future research and practice. By analysing the results from training the text-to-text generators we could establish a specific design brief that informed the final concept. The results from the image-to-image generator gave an overview of the shape grammars of previous submissions. All results were intriguing and can assist creativity and in this way, the tools were useful for gaining insight into historical architectural data, helped shape a specific design brief, and provoked new ideas. By reflecting on our design process, we argue that the use of language when employing such tools takes a new role and that three layers of language intertwined in our work: architectural discourse, programming languages, and annotations. We present a map that unfolds how these layers came together as a contribution to making machine learning more explainable for creatives. [ABSTRACT FROM AUTHOR]

Wind turbines (WTs) face a high risk of failure due to environmental factors like erosion, particularly in high-precipitation areas and offshore scenarios. In this paper we introduce a novel computational tool for the fast prediction of rain erosion damage on WT blades that is useful in operation and maintenance decision making tasks. The approach is as follows: Pseudo-Direct Numerical Simulation (P-DNS) simulations of the droplet-laden flow around the blade section profile are employed to build a high-fidelity data set of impact statistics for potential operating conditions. Using this database as training data, a machine learning-based surrogate model provides the feature of the impact pattern over the 2-D section for given wind and rain conditions. With this information, a fatigue-based model estimates the remaining lifetime and erosion damage for both homogeneous and coating-substrate blade materials. This prediction is done by quantifying the accumulated droplet impact energy and evaluating operative conditions over time periods for which the weather at the installation site is known. In this work, we describe the modules that compose the prediction method, namely the database creation, the training of the surrogate model and their coupling to build the prediction tool. Then, the method is applied to predict the remaining lifetime and erosion damage to the blade sections of a reference WT. To evaluate the reliability of the tool, several site locations (offshore, coastal, and inland), the coating material and the coating thickness of the blade are investigated. In few minutes we are able to estimate erosion after many years of operation. The results are in good agreement with field observations, showing the promise of the new rain erosion prediction approach. [ABSTRACT FROM AUTHOR]

This research evaluates the reliability of a machine vision system connected to a collaborative robot. In recent scientific papers, the authors have focused on machine vision itself, machine vision systems and related theory in general, along with machine learning methods, and image processing itself. However, there seems to be a missing link between these topics and the industrial robot's accuracy in basic tasks when utilizing machine vision. The experiments conducted, took place within an Industry 4.0 laboratory, where 3D-printed objects were utilized as test subjects. The collaborative robot, equipped with machine vision, performed tasks such as object removal and stacking. The evaluation focused on the success rate of object assembly and grasping. The paper discusses the integration of machine vision technology, previous research on reliability, and the use of a 2D camera for the collaborative robot. The findings contribute to understanding the potential of machine vision in enhancing efficiency and precision in collaborative robot workspaces. [ABSTRACT FROM AUTHOR]

A Mobile Ad-hoc Network (MANET) is a self-organizing collection of mobile devices communicating in a distributed fashion across numerous hops. MANETs are an appealing technology for many applications, including rescue operations, environmental monitoring, tactical operations, and so on, because they let people communicate without the usage of permanent infrastructure. This flexibility, however, creates additional security vulnerabilities. Because of its benefits and expanding demand, MANETs have attracted a lot of interest from the scientific community. They do, however, seem to be more vulnerable to numerous attacks that wreak havoc on their performance than any network. Traditional cryptography techniques cannot entirely defend MANETs in terms of fresh attacks and vulnerabilities due to the distributed architecture of MANETs; however, these issues can be overcome by using machine learning approaches-based intrusion detection systems (IDS). IDS, typically screening system processes and identifying intrusions, are commonly employed to supplement existing security methods because preventative techniques are never enough. Because MANETs are continually evolving, their highly limited nodes, and the lack of central observation stations, intrusion detection is a complex and tough process. Conventional IDSs are difficult to apply to them. Existing methodologies must be updated for MANETs or new approaches must be created. This paper aims to present a novel concept founded on deep belief networks (DBN) and long shortterm memory (LSTM) for MANET attack detection. The experimental analysis was performed on the probe, root to local, user to root, and denial of service (DoS) attacks. In the first phase of this paper, particle swarm optimization was used for feature selection, and subsequently, the DBN and LSTM were used for the classification of attacks in the MANET. The experimental results gave an accuracy reaching 99.46%, a sensitivity of 99.52%, and a recall of 99.52% for DBN and LSTM accuracy reaching 99.75%, a sensitivity of 99.79%, and a recall of 99.79%. [ABSTRACT FROM AUTHOR]

Reproducibility of artificial intelligence (AI) research has become a growing concern. One of the fundamental reasons is the lack of transparency in data, code, and model. In this work, we aimed to systematically review the radiology and nuclear medicine papers on AI in terms of transparency and open science. A systematic literature search was performed in PubMed to identify original research studies on AI. The search was restricted to studies published in Q1 and Q2 journals that are also indexed on the Web of Science. A random sampling of the literature was performed. Besides six baseline study characteristics, a total of five availability items were evaluated. Two groups of independent readers including eight readers participated in the study. Inter-rater agreement was analyzed. Disagreements were resolved with consensus. Following eligibility criteria, we included a final set of 194 papers. The raw data was available in about one-fifth of the papers (34/194; 18%). However, the authors made their private data available only in one paper (1/161; 1%). About one-tenth of the papers made their pre-modeling (25/194; 13%), modeling (28/194; 14%), or post-modeling files (15/194; 8%) available. Most of the papers (189/194; 97%) did not attempt to create a ready-to-use system for real-world usage. Data origin, use of deep learning, and external validation had statistically significantly different distributions. The use of private data alone was negatively associated with the availability of at least one item (p <0.001). Overall rates of availability for items were poor, leaving room for substantial improvement. [ABSTRACT FROM AUTHOR]

This paper proposes an industrial-friendly approach based on physics-informed Bayesian machine learning for probabilistic inference and refinement of stability predictions in milling. The knowledge from physics principles and theoretical models is gathered into a substructure-based framework and enables dedicated updating and exchange of models on a substructure level. The Bayesian learning algorithm utilizes this framework as a basis to simultaneously interact with multiple machine and process configurations that share substructures, leading to federated and transfer learning over a network of machine tools on a production site. Moreover, the paper further elaborates on the Bayesian perspective to anticipate the potential usefulness of given experimental data in improving the model accuracy by quantifying information gain for two primary purposes: Firstly, recursive online learning can manage computational resources by processing only the most informative data points that may be collected during arbitrary cuts. Secondly, active learning leverages the information gain to dynamically adapt a sequential data collection, leading to accelerated learning with minimal labeled data. The experimental validations confirm that the proposed approaches lead to effective learning and reliable predictions of milling stability, outperforming traditional deterministic methods. [ABSTRACT FROM AUTHOR]

Due to the anonymity of blockchain, frequent security incidents and attacks occur through it, among which the Ponzi scheme smart contract is a classic type of fraud resulting in huge economic losses. Machine learningbased methods are believed to be promising for detecting ethereum Ponzi schemes. However, there are still some flaws in current research, e.g., insufficient feature extraction of Ponzi scheme smart contracts, without considering class imbalance. In addition, there is room for improvement in detection precision. Aiming at the above problems, this paper proposes an ethereum Ponzi scheme detection scheme through opcode context analysis and adaptive boosting (AdaBoost) algorithm. Firstly, this paper uses the n-gram algorithm to extract more comprehensive contract opcode features and combine them with contract account features, which helps to improve the feature extraction effect. Meanwhile, adaptive synthetic sampling (ADASYN) is introduced to deal with class imbalanced data, and integrated with the Adaboost classifier. Finally, this paper uses the improved AdaBoost classifier for the identification of Ponzi scheme contracts. Experimentally, this paper tests our model in real-world smart contracts and compares it with representative methods in the aspect of F1-score and precision. Moreover, this article compares and discusses the state of art methods with our method in four aspects: data acquisition, data preprocessing, feature extraction, and classifier design. Both experiment and discussion validate the effectiveness of our model. [ABSTRACT FROM AUTHOR]

The two main causes of blindness are diabetes and glaucoma. Routine diagnosis of blindness is based on the conventional robust mass-screening method. However, despite being cost-effective, this method has some problems as a human eye-disease detection method because there are many types of eye disease that are similar or that result in no visual changes in the eye image. These issues make it highly difficult to recognize blindness and control it. Moreover, the color of the macula of the spot can be very close to that of the affected macula in a variety of eye diseases, which suggests that the color of the macula spot can indicate various possibilities, rather than one. This paper discusses the shortcomings of current blindness-screening and monitoring systems and presents a feature-based blindness diagnosis approach using digital eye fundus images for the purpose of automated diagnosis of eye disorders, considering three conditions: healthy eye, diabetic retinopathy (DR), and glaucoma. As such, this paper develops a computer-aided diagnosis (CAD) method for automated detection of human blindness. The proposed approach integrates Gabor filter features, statistical features, colored features, morphological features, and local binary pattern features, then compares them with features drawn from a standard dataset of 1580 fundus images. Several classification techniques were applied to the extracted-features neural network (NN), support vector machine (SVM), naïve bias (NB). SVM classifiers show the most promising accuracy. They achieved 93.3% over the other classifiers. [ABSTRACT FROM AUTHOR]

The proposed way of estimating the associated macroeconomic index based on a supply chain network is a novel strategy that might give useful insights for firms and organizations trying to enhance their supply chain logistics operations. To aid in this process, the suggested technique utilizes multiple regression analysis and adaptive extreme learning machine models to determine the relative importance of each indicator in the supply chain logistics decision-making process. Firms may be able to save expenses and boost economic value by employing enhanced genetic algorithms and mathematical modeling to develop a logistics distribution model for urban supply chains based on reconstructive genetic algorithms. It is possible that the paper's study of the issues plaguing the current distribution of logistics in metropolitan areas would prove valuable to organizations that are seeking to improve their own supply chain logistics procedures. This paper appears to take a thorough strategy that might assist assure the accuracy of the forecasting method by preprocessing macroeconomic indicators using imputation, classification, and other approaches to generate a time series consistency model. Some novel ways that may shed light on logistics in the supply chain include using a two-dimensional discrete mesh structure built on wireless sensors to depict the national economic development scenario and using a coding matrix to convey the extent of economic growth. A thorough and all-encompassing way that might aid organizations in making better judgments about their supply chain logistics strategies is to employ multiple regression analysis, an adaptive extreme learning machine model, and other approaches to examine the effect degree of each key indicator. Experiments demonstrating excellent and steady prediction accuracy of the algorithm model are encouraging and point to the possible usefulness of the suggested forecasting approach. The paper's contributions, including its analysis of supply chain logistics cost accounting, determination of the basic path optimization, improvement of the genetic algorithm, and design of the mathematical model of supply chain logistics distribution path decision, all look promising in terms of their potential to help businesses cut costs and boost economic value. The positive outcome of the successful design of the mathematical model demonstrates the possible efficacy of the suggested technique. [ABSTRACT FROM AUTHOR]

The rapid growth of Fintech innovation and the widespread adoption of blockchain technologies have indeed had a transformative impact on the financial industry. In this paper, the focus is on the application of machine learning algorithms, specifically the Random Forest Regression algorithm, within the context of Fintech and blockchain. This research contributes to the advancement of machine learning techniques in the field of Fintech and blockchain. The Random Forest Regression algorithm utilizes ensemble learning, combining multiple decision trees to analyze complex financial data and make predictions on various outcomes. This algorithm has proven to be effective in addressing key challenges within the industry, such as predicting loan defaults, detecting fraud, and assessing risks. Through experimental evaluations and case studies, the paper demonstrates the effectiveness of the Random Forest Regression algorithm in enhancing Fintech innovation in blockchain applications. The algorithm improved accuracy, scalability, and interpretability enable financial institutions to make data-driven decisions and optimize their operations. [ABSTRACT FROM AUTHOR]

Artificial intelligence and machine learning are revolutionising fields of science and engineering. In recent years, process engineering has widely benefited from this novel modelling and optimisation approach. The open literature can offer several examples of their applications to chemical engineering problems. Increasing investments are devoted to these techniques from different industrial areas, but insufficient information on a structured course covering these topics in a chemical engineering curriculum could be found. The course in this paper intends to reduce this gap. We introduce one of the first courses on artificial intelligence applications in a chemical engineering curriculum. The course targets Master's students with a chemical engineering background and insufficient knowledge of statistical approaches. It covers the main aspects by utilising frontal lectures and hands-on exercises with active learning methods. This paper shows the methodology we adapted to introduce students to machine learning techniques and how they responded to each class. The student performances for each test are shown, as well as the survey results based on student feedback and suggestions. This work contains essential guidelines for educators who will provide an artificial intelligence course in a chemical engineering curriculum. • An AI course for chemical engineers is given as one of the first attempts in Europe. • The course is structured into frontal and hands-on lessons. • Modelling and optimisation techniques via AI were shown to the students. • Applications of developed models were explained to the students. • Students can apply basic AI in chemical engineering tasks after this course. [ABSTRACT FROM AUTHOR]

Fuzzy rule-based systems (FRBSs) is a rule-based system which uses linguistic fuzzy variables as antecedents and consequent to represent human-understandable knowledge. They have been applied to various applications and areas throughout the soft computing literature. However, FRBSs suffers from many drawbacks such as uncertainty representation, high number of rules, interpretability loss, high computational time for learning. To overcome these issues with FRBSs, there exists many extensions of FRBSs. This paper presents an overview and literature review of recent trends on various types and prominent areas of fuzzy systems (FRBSs) namely genetic fuzzy system, hierarchical fuzzy system, neuro fuzzy system, evolving fuzzy system, FRBSs for big data, FRBSs for imbalanced data, interpretability in FRBSs and FRBSs which use cluster centroids as fuzzy rules. The review is for years 2010–2021. This paper also highlights important contributions, publication statistics and current trends in the field. The paper also addresses several open research areas which need further attention from the FRBSs research community. [ABSTRACT FROM AUTHOR]

To solve the problem of slow convergence and easy to get into the local optimum of the spider monkey optimization algorithm, this paper presents a new algorithm based on multi-strategy (ISMO). First, the initial population is generated by a refracted opposition-based learning strategy to enhance diversity and ergodicity. Second, this paper introduces a non-linear adaptive dynamic weight factor to improve convergence efficiency. Then, using the crisscross strategy, using the horizontal crossover to enhance the global search and vertical crossover to keep the diversity of the population to avoid being trapped in the local optimum. At last, we adopt a Gauss-Cauchy mutation strategy to improve the stability of the algorithm by mutation of the optimal individuals. Therefore, the application of ISMO is validated by ten benchmark functions and feature selection. It is proved that the proposed method can resolve the problem of feature selection. [ABSTRACT FROM AUTHOR]

The products of an archival culture in colleges and universities are the final result of the development of archival cultural resources, and the development of archival cultural effects in colleges and universities should be an important part of improving the artistic level of libraries. The existing RippleNet model doesn't consider the influence of key nodes on recommendation results, and the recommendation accuracy is not high. Therefore, based on the RippleNet model, this paper introduces the influence of complex network nodes into the model and puts forward the Cn RippleNet model. The performance of the model is verified by experiments, which provide a theoretical basis for the promotion and recommendation of its cultural products of universarchives, solve the problem that RippleNet doesn't consider the influence of key nodes on recommendation results, and improve the recommendation accuracy. This paper also combs the development course of archival cultural products in detail. Finally, based on the Cn-RippleNet model, the cultural effect of university archives is recommended and popularized. [ABSTRACT FROM AUTHOR]

For the assessment of optimum treatment timing in dentofacial orthopedics, understanding the growth process is of paramount importance. The evaluation of skeletal maturity based on study of the morphology of the cervical vertebrae has been devised to minimize radiation exposure of a patient due to hand wrist radiography. Cervical vertebral maturation assessment (CVMA) predictions have been examined in the state-of-the-art machine learning techniques in the recent past which require more attention and validation by clinicians and practitioners. This paper aimed to answer the question "How are machine learning techniques being employed in studies concerning cervical vertebral maturation assessment using lateral cephalograms?" A systematic search through the available literature was performed for this work based upon the Population, Intervention, Comparison and Outcome (PICO) framework. The searches were performed in Ovid Medline, Embase, PubMed and Cochrane Central Register of Controlled Trials (CENTRAL) and Cochrane Database of Systematic Reviews (CDSR). A search of the grey literature was also performed in Google Scholar and OpenGrey. We also did a hand-searching in the Angle Orthodontist, Journal of Orthodontics and Craniofacial Research, Progress in Orthodontics, and the American Journal of Orthodontics and Dentofacial Orthopedics. References from the included articles were also searched. A total of 25 papers which were assessed for full text, and 13 papers were included for the systematic review. The machine learning methods used were scrutinized according to their performance and comparison to human observers/experts. The accuracy of the models ranged between 60 and 90% or above, and satisfactory agreement and correlation with the human observers. Machine learning models can be used for detection and classification of the cervical vertebrae maturation. In this systematic review (SR), the studies were summarized in terms of ML techniques applied, sample data, age range of sample and conventional method for CVMA, which showed that further studies with a uniform distribution of samples equally in stages of maturation and according to the gender is required for better training of the models in order to generalize the outputs for prolific use to target population. [ABSTRACT FROM AUTHOR]

This paper is concerned with knowledge graph embedding with background knowledge, taking the formal perspective of logics. In knowledge graph embedding, knowledge--expressed as a set of triples of the form (a R b) ("a is R-related to b")--is embedded into a real-valued vector space. The embedding helps exploiting geometrical regularities of the space in order to tackle typical inductive tasks of machine learning such as link prediction. Recent embedding approaches also consider incorporating background knowledge, in which the intended meanings of the symbols a, R, b are further constrained via axioms of a theory. Of particular interest are theories expressed in a formal language with a neat semantics and a good balance between expressivity and feasibility. In that case, the knowledge graph together with the background can be considered to be an ontology. This paper develops a cone-based theory for embedding in order to advance the expressivity of the ontology: it works (at least) with ontologies expressed in the description logic ALC, which comprises restricted existential and universal quantifiers, as well as concept negation and concept disjunction. In order to align the classical Tarskian Style semantics for ALC with the sub-symbolic representation of triples, we use the notion of a geometric model of an ALC ontology and show, as one of our main results, that an ALC ontology is satisfiable in the classical sense iff it is satisfiable by a geometric model based on cones. The geometric model, if treated as a partial model, can even be chosen to be faithful, i.e., to reflect all and only the knowledge captured by the ontology. We introduce the class of axis-aligned cones and show that modulo simple geometric operations any distributive logic (such as ALC) interpreted over cones employs this class of cones. Cones are also attractive from a machine learning perspective on knowledge graph embeddings since they give rise to applying conic optimization techniques. [ABSTRACT FROM AUTHOR]

The financial time series data is a highly nonlinear signal and hence difficult to predict precisely. The prediction accuracy can be improved by linearizing the signal. In this paper the nonlinear data sample is linearized by decomposing it into several IMFs. A hybrid multi-layer decomposition technique is developed. The decomposition proposed in this paper is the combination of both EMD and VMD methods. As a new contribution to the previous literature in this study the VMD is used to further decompose the higher frequency signals obtained from the EMD based decomposed signal. In the result analysis it is observed that the double decomposition improves the prediction accuracy. This is a new introduction in the field of stock market prediction. The prediction accuracy of the proposed model is performed by applying it to three different stock markets for predicting the closing price. Historical data (closing price) is implemented to obtain 1 day ahead predicted closing price. Comparative analysis of different previously implemented methods like BPNN, SVM, ANN and ELM, along with the proposed method is performed. GA is implemented for optimizing the kernel factors. It is observed that the proposed hybrid model outperformed the other methods. [ABSTRACT FROM AUTHOR]

One of the most dangerous threats in WSNs is botnet attacks, in which attackers use mutual communications between IoT devices to launch large-scale malicious activities. In this regard, developments in the field of effective and reliable means of defence against this type of threat, in particular, reliable methods for detecting, identifying, and countering botnet attacks, are becoming increasingly important and relevant. This paper presents a comprehensive study that applies machine learning techniques, namely Random Forest and XGBoost, to identify botnet attacks on IoT effectively. These algorithms are analyzed, compared, and shown to be highly effective in detecting complex patterns indicative of botnet activity, thus achieving a significant improvement in IoT security. The conducted research aims to make a useful contribution to the problem of securing WSNs and IoT in general. The results of the study demonstrated high accuracy in detecting attacks with an accuracy of 99.18% for XGBoost and Random Forest showed an accuracy of 99.21%. Thus, it was shown that the significance of applying machine learning techniques such as Random Forest and XGBoost can be one of the key approaches in combating botnet attacks and securing the IoT. The results of the work emphasize the promising application of machine learning techniques for effective defense against cyber threats and highlight the importance of further. [ABSTRACT FROM AUTHOR]

This paper presents findings on machine learning agent behavior prediction in a board game application developed by a group of students. The goal of this research is to create a model facilitating collaboration between a user and an AI to play together in the board game using a Human-in-the-Loop architecture. By injecting explainability, the aim is to enhance communication and understanding between the user and the AI agent. Featuring a competitive Artificial Intelligence (AI) based on the Proximal Policy Optimization model, this research explores methods to make AI decisions transparent for enhanced player understanding. Two predictive models, a Decision Tree (DT) and a Deep Learning (DL) classifier, were developed and compared. The results show that the DT model is effective for short-term predictions but limited in broader applications, while the DL classifier shows potential for long-term prediction without requiring direct access to the game's AI. This study contributes to understanding human-AI interaction in gaming and offers insights into AI decision-making processes. [ABSTRACT FROM AUTHOR]

IoT technologies are becoming more and more common in our daily activities because the networks they create are capable of collecting information, monitoring and controlling remotely. However, these devices are not exempt from security attacks, as they become vulnerable entry points to data networks. The use of traditional methods to secure networks (e.g., Next Generation Firewalls (NGFW), encryption, etc.) is not recommended because the devices used in this type of network are limited in terms of computing power and storage availability (e.g., nodeMCU). In this paper, we propose to design two intrusion detection systems in embedded systems using machine learning (ML) algorithms, Artificial Neural Networks and K-means. In a distributed architecture Wireless Sensor Network scenario (WSN), we evaluate their performance in terms of connection and response times, detection accuracy and intruder detection time. Simulation results show that both models are able to find irregularities in network traffic within milliseconds. [ABSTRACT FROM AUTHOR]

Every data packet must pass through a few intermediate nodes to reach its destination. Among other reasons, tremendous growth in internet devices encourages those intermediate nodes to drop the data packets. Optimizing the data packet route is an effective solution to deal with packet loss. Advanced machine learning approaches have been identified as a powerful support tool for routing optimization in node networks. Cloud computing has kept pace with the continuously developing hardware infrastructure. Improved connection, processing power, and memory units enable real-time machine learning. This paper proposes and evaluates an approach for optimizing the packet path, by one hop, for intermediate nodes as a backup called Cloud Acknowledgement Scheme (CAS). It offers information on the transmission trend and the tendencies of certain adjacent nodes or groups of neighboring nodes in a network. The proposed CAS has been validated via a series of machine learning experiments using real-world node data. [ABSTRACT FROM AUTHOR]

Continual learning (CL) is a paradigm which addresses the issue of how to learn from sequentially arriving tasks. The goal of this paper is to introduce a CL framework which can both learn from a global multi-task architecture and locally adapt this learning to the task at hand. In addition to the global knowledge, we conjecture that it is also beneficial to further focus on the most relevant pieces of previous knowledge. Using a prototypical network as a proxy, the proposed framework bases its adaptation on the similarity between the current data stream and the previously encountered data. We develop two algorithms, one for the standard task-aware CL and another for the more challenging task-free setting where boundaries between tasks are unknown. We correspondingly derive a generalization upper bound on the error of an upcoming task. Experiments demonstrate that the introduced algorithms lead to improved performance on several CL benchmarks. [ABSTRACT FROM AUTHOR]

Chatbots are conversational software applications designed to interact dialectically with users for a plethora of different purposes. Surprisingly, these colloquial agents have only recently been coupled with computational models of arguments (i.e. computational argumentation), whose aim is to formalise, in a machine-readable format, the ordinary exchange of information that characterises human communications. Chatbots may employ argumentation with different degrees and in a variety of manners. The present survey sifts through the literature to review papers concerning this kind of argumentation-based bot, drawing conclusions about the benefits and drawbacks that this approach entails in comparison with standard chatbots, while also envisaging possible future development and integration with the Transformer-based architecture and state-of-the-art Large Language Models. [ABSTRACT FROM AUTHOR]

Decision-focused learning (DFL) is an emerging paradigm that integrates machine learning (ML) and constrained optimization to enhance decision quality by training ML models in an end-to-end system. This approach shows significant potential to revolutionize combinatorial decision-making in real-world applications that operate under uncertainty, where estimating unknown parameters within decision models is a major challenge. This paper presents a comprehensive review of DFL, providing an in-depth analysis of both gradient-based and gradient-free techniques used to combine ML and constrained optimization. It evaluates the strengths and limitations of these techniques and includes an extensive empirical evaluation of eleven methods across seven problems. The survey also offers insights into recent advancements and future research directions in DFL. [ABSTRACT FROM AUTHOR]

In recent years, there has been a proliferation of competing conceptions of what it means for a predictive algorithm to treat its subjects fairly. Most approaches focus on explicating a notion of group fairness, i.e. of what it means for an algorithm to treat one group unfairly in comparison to another. In contrast, Dwork et al. (2012) attempt to carve out a formalised conception of individual fairness, i.e. of what it means for an algorithm to treat an individual fairly or unfairly. In this paper, I demonstrate that the conception of individual fairness advocated by Dwork et al. is closely related to a criterion of group fairness, called ‘base rate tracking’, introduced in Eva (2022). I subsequently show that base rate tracking solves some fundamental conceptual problems associated with the Lipschitz criterion, before arguing that group level fairness criteria are at least as powerful as their individual level counterparts when it comes to diagnosing algorithmic bias. [ABSTRACT FROM AUTHOR]

Dyna-style reinforcement learning (RL) agents improve sample efficiency over model-free RL agents by updating the value function with simulated experience generated by an environment model. However, it is often difficult to learn accurate models of environment dynamics, and even small errors may result in failure of Dyna agents. In this paper, we highlight that one potential cause of that failure is bootstrapping off of the values of simulated states, and introduce a new Dyna algorithm to avoid this failure. We discuss a design space of Dyna algorithms, based on using successor or predecessor models—simulating forwards or backwards—and using one-step or multi-step updates. Three of the variants have been explored, but surprisingly the fourth variant has not: using predecessor models with multi-step updates. We present the Hallucinated Value Hypothesis (HVH): updating the values of real states towards values of simulated states can result in misleading action values which adversely affect the control policy. We discuss and evaluate all four variants of Dyna amongst which three update real states toward simulated states — so potentially toward hallucinated values — and our proposed approach, which does not. The experimental results provide evidence for the HVH, and suggest that using predecessor models with multi-step updates is a promising direction toward developing Dyna algorithms that are more robust to model error. [ABSTRACT FROM AUTHOR]

Anomaly detection (AD) task corresponds to identifying the true anomalies among a given set of data instances. AD algorithms score the data instances and produce a ranked list of candidate anomalies. The ranked list of anomalies is then analyzed by a human to discover the true anomalies. Ensemble of tree-based anomaly detectors trained in an unsupervised manner and scoring based on uniform weights for ensembles are shown to work well in practice. However, the manual process of analysis can be laborious for the human analyst when the number of false-positives is very high. Therefore, in many real-world AD applications including computer security and fraud prevention, the anomaly detector must be configurable by the human analyst to minimize the effort on false positives. One important way to configure the detector is by providing true labels (nominal or anomaly) for a few instances. Recent work on active anomaly discovery has shown that greedily querying the top-scoring instance and tuning the weights of ensemble detectors based on label feedback allows us to quickly discover true anomalies. This paper makes four main contributions to improve the state-of-the-art in anomaly discovery using tree-based ensembles. First, we provide an important insight that explains the practical successes of unsupervised tree-based ensembles and active learning based on greedy query selection strategy. We also present empirical results on real-world data to support our insights and theoretical analysis to support active learning. Second, we develop a novel batch active learning algorithm to improve the diversity of discovered anomalies based on a formalism called compact description to describe the discovered anomalies. Third, we develop a novel active learning algorithm to handle streaming data setting. We present a data drift detection algorithm that not only detects the drift robustly, but also allows us to take corrective actions to adapt the anomaly detector in a principled manner. Fourth, we present extensive experiments to evaluate our insights and our tree-based active anomaly discovery algorithms in both batch and streaming data settings. Our results show that active learning allows us to discover significantly more anomalies than state-of-the-art unsupervised baselines, our batch active learning algorithm discovers diverse anomalies, and our algorithms under the streaming-data setup are competitive with the batch setup. [ABSTRACT FROM AUTHOR]

Modeling policy effects in the context of high-dimensional data requires a balanced consideration of omitted interaction bias and overfitting problems. This paper investigates the role of machine learning algorithms in stabilizing estimates and demonstrates the possible regularization bias caused by common LASSO methods. To overcome the three problems simultaneously, postdouble selection is used to screen for the interaction terms that need to be included in the model, and the variance estimates are expanded to measure the uncertainty of the interaction effects and marginal effects. Monte Carlo simulations analyze the main factors affecting conditional and non-linear relationships: covariance and sample size. The results of empirical examples show that different model settings and estimation methods can lead to observable differences in the conclusion of treatment effect heterogeneity, and in general, post-double selection has better performance than other estimation methods. [ABSTRACT FROM AUTHOR]