Search

In the increasingly complex urban transportation landscape, the necessity for accurate, multi-step forecasting has never been more apparent to help plan for long-term, strategic transportation initiatives like Intelligent Transportation Systems. This work focuses on resolving uncertainties around the proper training step size for machine learning models performing multi-step traffic flow forecasting. Two strategies are considered: one-step and horizon-step training sizes, which extend from the single-step forecasting method. This paper compares and evaluates two machine learning models: the Support Vector Regression and Long Short-Term Memory (LSTM). Data from two country road locations, including urban roads in Kuala Lumpur and the I5-North freeway in California, are employed to forecast traffic flow in 1-hour increments, projecting from 2 h up to 24 h ahead. The results reveal a significant difference in performance between the two training step sizes, with the one-step training size emerging as the more consistent and optimal strategy for both Direct and Multi-Input Multi-Output forecasting strategies. The proposed model, the Direct Particle Swarm Optimization Grid Search Support Vector Regression (Direct-PSOGS-SVR) model, exhibited comparable or slightly better performance than the LSTM model in both environments. In the I5-North freeway dataset, the Direct-PSOGS-SVR achieved similar Root Mean Squared Error values across most forecasting tasks compared to the LSTM, indicating its effectiveness in stable traffic flow patterns. Notably, in the more dynamic urban environment of Kuala Lumpur, the Direct-PSOGS-SVR model also demonstrated stability and resilience in forecasting accuracy, effectively handling the inherent noise and fluctuations in traffic patterns. These findings serve as a valuable guide for practitioners in selecting the most efficacious combination of training strategies for specific time series forecasting tasks utilizing machine learning models, particularly in traffic flow forecasting. Introducing the Direct-PSOGS-SVR model enriches the landscape of machine learning solutions for traffic forecasting, underscoring the paper's contributions to the broader understanding of time series forecasting dynamics. [ABSTRACT FROM AUTHOR]

• We conducted a quantitative analysis of intelligent SCA technology over the past 20 years using bibliometric methods and data portrait techniques and traced earlier literature for a systematic review. • We identified the hotspots and research frontiers of intelligent ship collision avoidance in maritime transportation. • We discussed the contributions, innovations, and trends of future research in intelligent ship collision avoidance. With the increasing promotion of digital technology, safety issues faced by rush ships in the maritime industry have once again received attention. Ship collision avoidance (SCA) is not only a hot topic in the shipping industry but also an eye-catching issue in the development of intelligent ships. This paper provides a bibliometric and systematic overview of the literature on SCA to assist researchers in understanding the frontiers and recent trends of SCA. Based on the bibliographic portrait, a classification and grading study was conducted on the literature to provide a systematic review of it. A screening process was conducted on 851 relevant articles related to SCA and published in 2004–2023 in the Web of Science (WoS) database, and 526 highly relevant and high-quality papers were selected. Then, CiteSpace, VOSviewer software, and data visualization techniques were used to conduct a bibliographic portrait analysis on the selected papers. The evidence from these systematic literature reviews revealed the close collaboration relationships among researchers, research institutions, and countries or regions in the field of SCA studying shortly. Furthermore, the frontiers of the research on SCA were focused on three aspects, i.e. research subjects, technological methods, and novel algorithms in intelligent SCA. COLREG is a problem that must be considered in intelligent SCA. The future trends in intelligent SCA were explored and SCA technology was introduced from artificial intelligence (AI) so as to meet the safety requirement of maritime autonomous surface ships. AI algorithms such as machine learning and deep learning are key technologies for SCA. This research provides a theoretical basis and implementation directions of the research on SCA. The hybrid encounters between traditional ships and intelligent ships, as well as multi-ship encounters in narrow waterways, will be the focal points of attention in the future. [ABSTRACT FROM AUTHOR]

Machine learning methods are widely used in Android malware detection, but carefully scrambled adversarial samples can effectively evade detection. Adversarial defence methods such as retraining and ensemble learning have high training costs and significant errors. This paper proposes a novel Android malware detection system, KFFPDet, to detect adversarial samples. It can detect adversarial samples efficiently. Firstly, this paper uses K-Nearest Neighbor (KNN) as the primary detection algorithm. In addition, this paper proposes a three-stage adversarial sample detection method, which saves the detection cost and reduces the error through prior knowledge. In the first stage, the frequency difference between benign typical features and malicious typical features is used to filter adversarial samples. In the second stage, the feature continuous values of benign typical features are used for adversarial sample filtering. The third stage filters adversarial samples by detecting isolation permissions and application programming interface (API). Experiments on various benchmark datasets verify the detection effect of KFFPDet. KFFPDet achieves 96% accuracy of adversarial sample detection on the VirusShare dataset. Compared with other advanced methods, KFFPDet 's adversarial sample detection accuracy is improved by 4.43% to 94.53%. The experimental results show that the proposed three-stage adversarial sample detection method is useful and will help design a more effective adversarial sample detection scheme. [ABSTRACT FROM AUTHOR]

• Gas turbine life management monitors and assesses turbine status and conditions. • Inlet and exit air passages impact gas turbine performance, not physical attributes. • Study gap: An incomplete gas-turbine model limits big data's predictive analysis. • Big data analytics drive optimized informed integrated health monitoring decisions for gas turbines. • Data mining extends gas turbine life by forecasting faults for the future. In addition to power generation, gas turbines (GT) are crucial components in many engineering applications, including those for aircraft, ships, and other machinery that needs a prime mover. One of the main concerns was predicting a comprehensive GT model, including intake, exit, and power-generator characteristics. Hence, this study provides a comprehensive review of GT's main components, parameters, faults, and predicting models using machine learning and data mining applications. Forty-four publications between 2014 and 2023 were chosen for this review study based on the primary keywords associated with the paper title. This study focuses solely on the GT model prediction aspects, considering the significance of field distribution, applications of data mining, research goals, benefits, constraints, and contributions to aviation, the oil and gas industry, and marine propulsion fields. Based on the examined studies, ANN, ELM, and KF applications show promise for potential GT life management. The typical limitations of GT life prediction, as observed in 93.2% of the review papers, involved the absence of a comprehensive GT model that incorporates all vibration factors and the use of substantial amounts of real data for predictive modelling. Another notable finding that underscores the knowledge gap in this area is the utilization of complete GT units to monitor unit life and facilitate life management. Apparently, the development of GT life management is still riddled with several limitations, which call for more improvement. [ABSTRACT FROM AUTHOR]

Multiple Instance Learning (MIL) is a learning paradigm, where training instances are arranged in sets, called bags, and only bag-level labels are available during training. This learning paradigm has been successfully applied in various real-world scenarios, including medical image analysis, object detection, image classification, drug activity prediction, and many others. This survey paper presents a comprehensive analysis of MIL, highlighting its significance, recent advancements, methodologies, applications, and evolving trends across diverse domains. The survey begins by explaining the core principles that form the basis of MIL and how it differs from traditional learning approaches. This sets the foundation for comprehending the distinct challenges and techniques of solving MIL problems. Next, we discuss how supervised learning algorithms are tailored to support MIL and combine this discussion with a review of seminal MIL algorithms as well as the latest innovations that incorporate neural networks, deep learning architectures, and attention techniques. This comprehensive analysis helps to understand the strengths, limitations, and adaptability of these methods across diverse data modalities, complexities, and applications. In summary, this survey paper provides an essential resource for researchers, practitioners, and enthusiasts seeking a comprehensive understanding of Multiple Instance Learning. It covers foundational concepts, traditional methods, recent advancements, and future directions. By providing a holistic view of MIL's dynamic landscape, this paper aims to inspire further innovation and exploration in this ever-evolving field. • A survey on the current state of the Multiple Instance Learning. • We provide applications of Multiple instance learning in various domains. • We discuss how existing supervised learning algorithms are modified to support MIL. • Publicly available datasets and open research challenges in MIL are discussed. [ABSTRACT FROM AUTHOR]

This paper proposes physics-informed meta-learning-based surrogate modeling (PI-MLSM), a novel approach that combines meta-learning and physics-informed deep learning to train surrogate models with limited labeled data. PI-MLSM consists of two stages: meta-learning and physics-informed task adaptation. The proposed approach is demonstrated to outperform other methods in four numerical examples while reducing errors in prediction and reliability analysis, exhibiting robustness, and requiring less labeled data during optimization. Moreover, compared to other approaches, the proposed approach exhibits better performance in solving out-of-distribution tasks. Although this paper acknowledges certain limitations and challenges, such as the subjective nature of physical information, it highlights the key contributions of PI-MLSM, including its effectiveness in solving a wide range of tasks and its ability in handling situations wherein physical laws are not explicitly known. Overall, PI-MLSM demonstrates potential as a powerful and versatile approach for surrogate modeling. • A proposed approach uses meta-learning and PIDL to train surrogates with small data. • A model-agnostic meta-learning model is introduced to learn surrogate model weights. • A PIDL approach is introduced to adapt to target tasks with meta-learned weights. • Our approach outperformed other methods in numerical examples given limited data. • Proposed approach showed robustness in solving out-of-distribution tasks. [ABSTRACT FROM AUTHOR]

Automatic humor detection has compelling use cases in modern technologies, such as humanoid robots, chatbots, and virtual assistants. In this paper, we propose a novel approach for detecting and rating humor in short texts based on a popular linguistic theory of humor. The proposed technical method initiates by separating sentences of the given text and utilizing the BERT model to generate embeddings for each one. The embeddings are fed to a neural network as parallel lines of hidden layers in order to determine the congruity and other latent relationships between the sentences, and eventually, predict humor in the text. We accompany the paper with a novel dataset consisting of 200,000 short texts, labeled for the binary task of humor detection. In addition to evaluating our work on the novel dataset, we participated in a live machine-learning competition to rate humor in Spanish tweets. The proposed model obtained F1 scores of 0.982 and 0.869 in the performed experiments which outperform general and state-of-the-art models. The evaluation results confirm the model's strength and robustness and suggest two important factors in achieving high accuracy in the current task: (1) usage of sentence embeddings and (2) utilizing the linguistic structure of humor in designing the proposed model. • A novel method for humor detection and rating based on a general linguistic theory of humor. • Introduced a very large novel dataset with 200k short texts for humor detection. • Achieved 98% accuracy and outperformed five strong baselines on the new dataset. • Demonstrated accuracy and robustness in a data science competition for Spanish texts. • The novel approach and dataset contributed to tens of research projects. [ABSTRACT FROM AUTHOR]

• Treat a problem of the imbalanced customer churn classification. • A new multi-strategy collaborative processing method IADASYN-FLCatBoost is proposed. • The traditional ADASYN algorithm is improved and an IADASYN algorithm is proposed. • The Focal Loss function is embedded into the CatBoost to form a new FLCatBoost. • Empirical analysis shows the proposed method performs better than similar methods. The rapid advancement of big data and artificial intelligence heralds a dual-edged era of opportunities and challenges for the banking sector. Indeed, enhancing a model's capability to accurately classify imbalanced datasets represents a critical challenge within the field of customer churn prediction (CCP). In this paper, to address the challenges presented by the problem of imbalanced customer classification, a new multi-strategy collaborative processing method named IADASYN-FLCatBoost is proposed from dual perspectives: data and algorithm. At the data level, the traditional Adaptive Synthetic (ADASYN) sampling is improved, that is, the LOF (Local Outlier Factor) algorithm is introduced to eliminate outliers, and the classification features are specially processed to synthesize new minority class samples, thus an improved ADASYN (IADASYN) algorithm is obtained. At the algorithm level, the Focal Loss is embedded into the CatBoost ensemble learning framework to form a new Focal Loss-CatBoost (FLCatBoost) to make a focal-aware, cost-sensitive version of imbalanced customer churn prediction. Moreover, the empirical analysis is conducted in conjunction with the credit card customer dataset obtained from the Kaggle platform. The results of the staged comparison experiments show that the proposed method IADASYN-FLCatBoost in this paper shows the best prediction performance. Comparing the proposed method with 5 other imbalanced classification algorithms and 20 classifiers composed of classical sampling methods and ensemble learning algorithms, it is verified that the classification effect of the proposed method performs best, and the values of Recall , F 1 score, G-mean and Area under Precision-Recall curve (AUPRC) have been significantly improved. In addition, further verification of the model also proves that the proposed method has certain generalizability and is still valid for other banks and customer churn datasets of other industries. [ABSTRACT FROM AUTHOR]

• Newbie study using standard ML techniques with HAR Application and Discussions. • Activities found in Literature with the corresponding reference. • Co-occurrences between activities and sensors with the corresponding reference. • Summary and comparison among the different datasets found in Literature. • Summary of the experimentation settings with performance scores found in Literature. Human Activity Recognition (HAR) is an essential area of research related to the ability of smartphones to retrieve information through embedded sensors and recognize the activity that humans are performing. Researchers have recognized people's activities by processing the data received from the sensors with Machine Learning Models. This work is intended to be a hands-on survey with practical's tables capable of guiding the reader through the sensors used in modern smartphones and highly cited developed machine learning models that perform human activity recognition. Several papers in the literature have been studied, paying attention to the preprocessing, feature extraction, feature selection, and classification techniques of the HAR system. In addition, several summary tables illustrating HAR approaches have been provided: most popular human activities in the literature with paper references, the most popular datasets available for download (Analyzing their characteristics, such as the number of subjects involved, the activities recorded, and the sensors with online-availability), co-occurrences between activities and sensors, and a summary table showing the performance obtained by researchers. =The paper's goal is to recommend, through the discussion phase and thanks to the tables, the current state of the art on this topic. [ABSTRACT FROM AUTHOR]

Roundwood sorting is still a manual process in many Swiss sawmills, requiring employees to visually inspect and categorize thousands of logs per day. The heavy workload can be both physically and mentally taxing and can lead to increased rates of human error. State-of-the-art automation systems like X-ray log scanners are expensive and difficult to integrate into existing process lines. This paper proposes a novel recommendation system that leverages recent advances in image classification to automate roundwood classification by quality and species. The system integrates a camera to capture cross-sectional images of logs and record numerical data, such as length, taper, and diameter. The analysis of the resulting dataset highlights the challenges of data imbalance and noise, which makes classification difficult and, in some cases, impossible. However, by using selected datasets with reduced noise, state-of-the-art Convolutional Neural Networks (CNNs) can extract quality and species features. Quality models learn from a manually selected and simplified dataset, featuring samples that experts can clearly classify based on the image's information. Species models are trained on a label-noise-reduced dataset, reflecting real-world complexity. The accuracy on the selected dataset for three quality classes is 80%. The species determination is less challenging and reaches 91% accuracy on a synchronized dataset for the main species spruce and fir. Overall, this paper highlights the potential of Machine Learning in augmenting the roundwood sorting processes and presents a novel system that can improve the efficiency and accuracy of the process. [Display omitted] • Automation of roundwood sorting: Replaces manual sorting with image-based AI. • Integrated camera system in roundwood sorting to collect labeled dataset. • Species prediction: 91% accuracy in spruce–fir distinction. • Quality prediction on complexity reduced dataset: 80% accuracy between three main quality levels. • Efficient, adaptable & scalable system which is easy to integrate into existing process lines. [ABSTRACT FROM AUTHOR]

The emergence of online social networks (OSN) platforms removes communication barriers that are essential to human life, catalyzing social networking growth. However, this emergence has given rise to a negative impact when someone abuses the platform to commit cyberbullying activities. Hence, it is crucial to work on automated cyberbullying-related classification to mitigate the societal phenomena in OSN. The research on the automated classification model for cyberbullying was pioneered over the last decade with growing interest among researchers. It is helpful to track its growth over the decades to elucidate the state-of-arts techniques applied in this field. This paper presents a large amount of literature germane to cyberbullying classification from past to present to provide a comprehensive review. A total of 126 papers were reviewed. This paper emphasizes text-based cyberbullying and multi-modal cyberbullying. The review was presented around the machine learning workflow, encompassing four core sections: dataset analysis, pre-processing analysis, feature analysis, and technique analysis. Based on the critical analysis, limitations are addressed along with the future works that can be conducted to fill the gap in previous research. Furthermore, the review also examined the ethical implications associated with the implementation of these techniques. This review paper is expected to assist readers in fully comprehending the current trend, architecture, and techniques applied to the field. [ABSTRACT FROM AUTHOR]

Truck drivers are essential to modern supply chains. However, the companies that employ them are having trouble keeping them. At many trucking firms, annual turnover rates approach 100 percent or higher. This paper develops a method for predicting individual truck driver turnover events before they happen by applying supervised machine learning classifiers to a new source of operational truck driver data. Our paper uses Electronic Logging Device (ELD) data, which comprise newly federally-mandated, time-stamped work logs collected from approximately 1200 American long-haul truck drivers over 3 years. We train three supervised machine learning classifiers (logistic regression, random forests, and support vector machines) on this data and achieve 60 to 70 percent prediction accuracy and 50 to 60 percent recall across two 5-fold cross-validated experiments. We observe that the quantity and consistency of week-day driving assignments explain most of our models' predictive power. We offer these results as both a new technical tool, as well as novel managerial insights for improving global truck driver retention problems for the benefit of supply chains worldwide. [Display omitted] • Newly available operational data from long haul truck drivers is analyzed. • Three machine learning classifiers are trained on the operational data. • Predictive models achieve 60–70 percent accuracy, 50–60 percent recall. • Models show potential to enable interventions to retain truck drivers. [ABSTRACT FROM AUTHOR]

Ignoring the temporal relationship of position points in trajectory data or the change in the number of position points in dummy trajectory generation can result in inadequate data availability. To address the data availability problem of publishing user trajectory data while ensuring privacy protection, this paper proposes an improved privacy protection method combining a Bidirectional Gated Recurrent Unit and differential privacy (BiGRU-DP). First, Laplace noise is added to the time attribute of the trajectory data. Secondly, BiGRU is used to predict the processing of trajectory data, considering the bi-directional data of the predicted point. Then, the k -means clustering method is used to generalize and optimize the trajectory data. Finally, the trajectory data is published after adding noise. The experimental simulation results show that BiGRU-DP enhances the availability of published data and has certain efficiency advantages over traditional trajectory data publishing methods while ensuring privacy protection. • Propose a trajectory data temporal attribute noise addition strategy. • Design a trajectory prediction algorithm to yield a more accurate trajectory. • Design a k -means trajectory timestamp position generalization processing algorithm. • Propose a trajectory exception-handling mechanism and post-processing mechanism. [ABSTRACT FROM AUTHOR]

Computational ecoacoustics has seen significant growth in recent decades, facilitated by the reduced costs of digital sound recording devices and data storage. This progress has enabled the continuous monitoring of vocal fauna through Passive Acoustic Monitoring (PAM), a technique used to record and analyse environmental sounds to study animal behaviours and their habitats. While the collection of ecoacoustic data has become more accessible, the effective analysis of this information to understand animal behaviours and monitor populations remains a major challenge. This survey paper presents the state-of-the-art ecoacoustics data analysis approaches, with a focus on their applicability to large-scale PAM. We emphasise the importance of large-scale PAM, as it enables extensive geographical coverage and continuous monitoring, crucial for comprehensive biodiversity assessment and understanding ecological dynamics over wide areas and diverse habitats. This large-scale approach is particularly vital in the face of rapid environmental changes, as it provides crucial insights into the effects of these changes on a broad array of species and ecosystems. As such, we outline the most challenging large-scale ecoacoustics data analysis tasks, including pre-processing, visualisation, data labelling, detection, and classification. Each is evaluated according to its strengths, weaknesses and overall suitability to large-scale PAM, and recommendations are made for future research directions. • Passive Acoustic Monitoring (PAM) enables continuous species monitoring. • We evaluate ecoacoustics datasets for applicability in large-scale PAM. • We compare key ecoacoustics data analysis frameworks/methods. • We discuss the open challenges/opportunities of large-scale PAM data analysis. [ABSTRACT FROM AUTHOR]

Acoustic emission (AE) is defined as the structure-borne propagating transient elastic stress waves produced due to the rapid release of energy because of internal structure changes or asperities contact under loading. AE sensor is one of the promising sensing technologies for machinery and structures health monitoring. A start-to-end comprehensive condition monitoring solution utilizes the potential of four technologies, i.e., platform, data, analytics, and operations. This study reviews these four technologies developed based on AE sensors for rotating machinery element's fault diagnosis and prognosis. Operating parameters influence the AE signal; hence, the influence of these parameters on AE characteristics is also discussed. Based on this review, challenges and future directions while using the AE sensor are detailed. No dedicated paper that summarises all these technologies in the context of AE sensors is available. Hence, using AE signals, analysis and measurements are mostly done by researchers based on their personal experiences. This article attempts to summarise the standard practices used by various researchers for AE data measurement and analysis for rotating machinery condition monitoring. [ABSTRACT FROM AUTHOR]

Predictive analytics aims to build machine learning models to predict behavior patterns and use predictions to guide decision-making. Predictive analytics is human involved, thus the machine learning model is preferred to be interpretable. In literature, Generalized Additive Model (GAM) is a standard for interpretability. However, due to the one-to-many and many-to-one phenomena which appear commonly in real-world scenarios, existing GAMs have limitations to serve predictive analytics in terms of both accuracy and training efficiency. In this paper, we propose FXAM (Fast and eXplainable Additive Model), a unified and fast interpretable model for predictive analytics. FXAM extends GAM's modeling capability with a unified additive model for numerical, categorical, and temporal features. FXAM conducts a novel training procedure called Three-Stage Iteration (TSI). TSI corresponds to learning over numerical, categorical, and temporal features respectively. Each stage learns a local optimum by fixing the parameters of other stages. We design joint learning over categorical features and partial learning over temporal features to achieve high accuracy and training efficiency. We prove that TSI is guaranteed to converge to the global optimum. We further propose a set of optimization techniques to speed up FXAM's training algorithm to meet the needs of interactive analysis. Thorough evaluations conducted on diverse data sets verify that FXAM significantly outperforms existing GAMs in terms of training speed, and modeling categorical and temporal features. In terms of interpretability, we compare FXAM with the typical post-hoc approach XGBoost+SHAP on two real-world scenarios, which shows the superiority of FXAM's inherent interpretability for predictive analytics. • A unified and fast interpretable model (FXAM) for predictive analytics is proposed. • FXAM extends modeling capability and efficiency of Generalized Additive Model. • FXAM conducts a novel training procedure called Three-Stage Iteration. • Joint learning and partial learning help achieve high accuracy and efficiency. • FXAM performs much better on synthetic data sets and 13 real datasets. [ABSTRACT FROM AUTHOR]

Although the accurate potential for growth prediction is very important for Government grants and contributions programs to better support Small and Medium-sized Enterprises (SMEs), it is a challenging task due to the data heterogeneity (both structured data and free text data bilingual in English and French), the class imbalance issue, and the difficulties in efficient feature learning. To address these challenges, this paper presents a novel BERT-TCN model for portfolio predictions in government funding programs, with the following key contributions. First, we describe the application of a novel architecture to a prediction task involving sequential, structured, partially quantitative input data and free text input data. Specifically, our novel model predicts the growth of firms receiving government funding for innovation. Our model also deals with class imbalance in the data and the difficulties in efficient feature learning. Our model integrates a Transformer model, i.e., BERT, for text modeling with a Temporal Convolutional Network (TCN) for sequential prediction. Second, we also developed various performance evaluation criteria in Section 4.3, allowing comprehensive assessments of the proposed approach from both the machine learning perspective and funding program-specific perspective. Third, the importance of features (both text and numerical features) is quantified and evaluated, allowing insights into how different features contribute to the prediction and explainability of the proposed model. The proposed approach is trained and tested on a large dataset from a rich database, demonstrating that the proposed approach can greatly help individual human experts improve their results. • Proposing a novel BERT-TCN model for text modeling and temporal prediction. • Designing various evaluation criteria for comprehensive assessments of the model. • Exploring explainability of the model by ablation studies of various input features. [ABSTRACT FROM AUTHOR]

When engaging in activities such as using video display terminals, driving, or sports, effective fatigue monitoring is crucial. Nevertheless, obtaining biological information through contact devices may interfere with normal activities. Deterministic expressions in mainstream machine learning have limitations in reflecting the continuous, dynamic, and fuzzy process of fatigue status. Additionally, fatigue is a cumulative process where the previous state impacts the assessment of the current one. To address these needs and challenges, this paper proposes a novel model based on facial videos called the temporal adaptive fuzzy neural network (TAFNN) for fatigue assessment. TAFNN utilizes an adaptive fuzzy neural network as its foundation and employs causal and dilated convolutions for the time information processing method to achieve time series extraction and fatigue assessment. It leverages facial physiological and motion features to minimize interference during assessment. Furthermore, TAFNN introduces a new calculation method for rule antecedents to enhance stability. Experimental results demonstrate that TAFNN effectively captures the cumulation and fuzziness of state changes, outperforming other widely adopted methods in both assessment ability and runtime performance. The improved rule antecedent calculation method successfully mitigates the issue of multiple memberships rapidly approaching zero after combination. Through 1000 repeated experiments, the enhanced method reduces TAFNN's instability by 81.58%. • A novel temporal adaptive fuzzy neural network (TAFNN) for assessment is introduced. • TAFNN can simultaneously express fatigue state changes' cumulation and fuzziness. • The improved rule antecedent calculation method can enhance the stability of TAFNN. • TAFNN has advantages in both fatigue assessment capability and runtime performance. [ABSTRACT FROM AUTHOR]

The advancement of machine learning and symbolic approaches have underscored their strengths and weaknesses in Natural Language Processing (NLP). While machine learning approaches are powerful in identifying patterns in data, they often fall short in learning commonsense and the factual knowledge required for the NLP tasks. Meanwhile, the symbolic methods excel in representing knowledge-rich data. However, they struggle to adapt dynamic data and generalize the knowledge. Bridging these two paradigms through hybrid approaches enables the alleviation of weaknesses in both while preserving their strengths. Recent studies extol the virtues of this union, showcasing promising results in a wide range of NLP tasks. In this paper, we present an overview of hybrid approaches used for NLP. Specifically, we delve into the state-of-the-art hybrid approaches used for a broad spectrum of NLP tasks requiring natural language understanding, generation, and reasoning. Furthermore, we discuss the existing resources available for hybrid approaches for NLP along with the challenges and future directions, offering a roadmap for future research avenues. • Review of hybridization of machine learning and symbolic methods for NLP. • Taxonomy of NLP tasks with substantial use of hybrid approaches. • Overview of state-of-the-art hybrid approaches for NLU, NLG, and NLR tasks. • Open challenges and future directions of Hybrid NLP. [ABSTRACT FROM AUTHOR]

A new model of utility analysis and performance prediction in crowdfunding is proposed in this paper with considering backer's behavior-related decision. To establish a connection between the backer's behavior-related decision and crowdfunding performance, the proposed model includes a modeling analysis and an empirical analysis. The modeling analysis focuses on analyzing backer's backing decision with establishing the utility-driven model, while the empirical analysis verifies the association between the utility-related features and crowdfunding performance, as well as the prediction improvement of the proposed model. Specifically, prospect theory (PT) is applied first in the modeling analysis part to analyze backer's prospect utilities, and then evidential theory (ET) is applied to aggregate the multi-source dynamic feature's utility. To validate the proposed model, and based on the results of modeling analysis, we conduct the significant test with regression model and the prediction test with several machine-learning prediction model to verify whether the proposed utility features can improve the prediction results in terms of crowdfunding performance. The empirical results show that the correlation between behavior-related factors and crowdfunding performance is significant. And the area under the curve (AUC) index of the proposed model is, on average, 7.39% higher than the baseline model on several machine-learning prediction models. In addition to providing new insights for predicting crowdfunding results, this research provides a powerful tool to assist crowdfunding platform and fundraisers in operational management with analyzing backers' backing behavior. • Modeling backers' behavior-related decisions in crowdfunding. • Empirical results validate the effect of behavior-related features on performance. • On average, the proposed model improves prediction accuracy by 7.39%. [ABSTRACT FROM AUTHOR]

A missing data problem is inevitable when collecting time series datasets from marine sensors. Due to this, sensor data is not reliable enough to assist decision-making. To impute missing values, a number of methods are available. Choosing the best imputation method, however, is not a trivial task, as it usually involves domain expertise and trial-and-error iterations. Additionally, if imputations are done carelessly, they produce a high error, resulting in incorrect assumptions by stakeholders. In this paper, a meta-learning approach is presented that can be used to extract characteristics of the underlying data, and based on that, a less error-prone imputation method is recommended. Ten commercial ocean-going vessel datasets are used to evaluate our proposed method. A total of 29,527 data samples were generated, comprising 22 inputs and 1 target. The proposed method achieves a weighted F1-Score of 87.5% when utilizing stratified 10-fold cross-validation. Our approach can improve the average imputation score up to 86%, with the worst-case improvement being 5%. This demonstrates that our proposed approach is efficient and effective in recommending the best imputation methods. • The characteristics of the dataset influence the recommendation of imputation methods. • Meta-learning can model the relationship between dataset characteristics and the performance of imputation methods. • Interpolation and MissForest are the most effective imputation techniques. [ABSTRACT FROM AUTHOR]

Machine learning methods have shown great performance in many areas, including neuroimaging data analysis. However, model performance is only one objective in neuroimaging analysis. Gaining insight from the data is also critical in this field, such as identifying regions where detected signals are relevant to cognitive and diagnostic tasks. To fulfill this need, enabling the explainability of a model's decision-making process is critical. Predictions of complex machine learning models are notoriously difficult to explain. This limits the use of complex models like kernel support vector machines (SVM) in neuroimaging analysis. Recently, several permutation-based methods have been developed to explain these complex models. However, the explanation results are affected by class-irrelevant features like suppressor variables and high background noise variables. This problem may also happen when explaining linear models. One possible reason is that the permutation process will produce unrealistic data instances when features are not independent, e.g. correlated. These unrealistic data instances will influence the explanation results. In neuroimaging analysis, the activation pattern, the estimated weight of the assumed generative model corresponding to the current classifier, is used to deal with this problem for linear models. This method does not rely on a permutation process but rather on the available data information. In this paper, we propose a novel method of Explanation through Activation Pattern (EAP) to explain the SVM models with different types of kernels for neuroimaging data analysis. Our method can generate a global feature importance score by estimating the activation pattern of kernel SVM models. We evaluate our method against three popular methods on both simulation datasets and a publicly available EEG/MEG dataset on visual tasks. The experimental results demonstrate that the proposed EAP method can provide explanations with low computational cost and is less affected by class-irrelevant features than the other three methods. In the experiment using the MEG/EEG dataset of visual tasks, the proposed EAP method provides agreement results with the brain's electrical activity patterns reported in the literature on the visual tasks EEG/MEG data and is significantly faster than the other explanation methods. • Explaining nonlinear SVM model predictions using activation pattern. • Explanation results robust to noise variables. • Significantly faster explanation than other state-of-art explanation methods. • Generating global feature importance scores. [ABSTRACT FROM AUTHOR]

Feature selection refers to the process of identifying the most relevant and valuable features or characteristics in a dataset, and it is an important step that can improve the performance and accuracy of predictive models for machine learning and data mining. Feature selection is crucial in efficiently reducing the number of features, thereby enhancing classification accuracy. This, in turn, alleviates computational burdens and elevates the overall performance of machine learning algorithms. This paper extended the conventional manta ray foraging optimization (MRFO) algorithm to an improved binary MRFO (IBMFO) algorithm. IBMRFO serves as a search strategy for designing a wrapper-based feature selection approach. To commence, we introduce a novel mechanism, termed the chaotic tent map (CTM), aimed at initializing the population to fortify both the algorithm's exploitation capabilities and population diversity. Second, an adaptive somersault foraging (ASF) factor is tuned to avoid premature convergence. Finally, a binary mechanism is introduced to adapt the algorithm for binary feature selection dilemmas. Experiments are performed on 30 renowned UC Irvine machine learning repository datasets. The results of the student's t-test show that the classification algorithm has statistical significance. The findings unequivocally exhibit the superior performance of the proposed IBMRFO compared to other algorithms under scrutiny. Notably, the efficiency of the enhanced factors is assessed through rigorous testing. For post-publications supporting this work, readers can refer to https://github.com/ELF-CHEUNG/IBMRFO. • Novel chaotic tent map for population initialization from evolutionary computation. • Optimization methods based on enhanced evolutionary computation. • Adaptive factor enhances convergence rate and search efficiency of algorithm. • Extensive numerical and scientific simulations for verifications with UCI dataset. [ABSTRACT FROM AUTHOR]

Dynamic events and transportation constraints would significantly affect the full utilization of resources and the reduction of production costs in distributed job shops. Therefore, in this paper, a deep reinforcement learning algorithm (DRL)-based real-time scheduling method is developed to minimize the mean tardiness of the dynamic distributed job shop scheduling problem with transfers (DDJSPT) considering random job arrivals. Firstly, the proposed DDJSPT is modeled as a Markov decision process (MDP). Then, ten problem-oriented state features covering four aspects of factories, machines, jobs, and operations are elaborately extracted from the dynamic distributed job shop. After that, eleven composite rules considering the uniqueness of DDJSPT are constructed as a pool of actions to intelligently prioritize unfinished jobs and allocate the selected job to an appropriate factory. Moreover, a justified reward function adapted from the objective is designed for better convergence of DRLs. Subsequently, five DRLs are employed to address the DDJSPT, encompassing deep Q-network (DQN), double DQN (DDQN), dueling DQN (DlDQN), trust region policy optimization (TRPO), and proximal policy optimization (PPO). Finally, grounded in numerical comparison experiments under 243 production configurations of the DDJSPT, the effectiveness and generalization of DRL-based scheduling methods are credibly verified and confirmed. [ABSTRACT FROM AUTHOR]

Band selection is one of the most important considerations for hyperspectral imaging (HSI) datasets that involves selecting fewer bands with higher classification accuracy. As a combinatorial optimization problem, evolutionary algorithm has been applied in the field, but the optimization ability and number of selected bands are limited with the control of exploration and exploitation. In this paper, an adaptive evolutionary-reinforcement learning algorithm (ERLA) is proposed for band selection of HSI datasets, only a suitable phase contributes to updating the corresponding agents in each iteration, and the coding length is adaptively lessened with the potential solution space. In addition, reinforcement learning is used to choose a feasible action in each iteration, and the unselected bands are removed from the candidate band combination in the previous state. A simulation study is conducted on four public HSI datasets, where the proposed framework has delivered an overall accuracy of 98% for 12% of bands in Chikusei dataset, 92% for 11% in KSC dataset, 93% against 19% for Xiongan dataset, and 97% in Longkou dataset with 10% bands retained, demonstrating that the proposed framework outperforms other approaches with several targets, and that the number of selected bands is obviously decreased with satisfactory classification accuracy. • A new ERLA is proposed for band selection that integrates the pattern of RL with EA. • A band selection framework is designed based on the characteristics of HSI datasets. • An effective reward mechanism is developed to evaluate the subsequent agents. • Less than 20% of bands is remained for each HSI dataset with a high enough OA. [ABSTRACT FROM AUTHOR]

• One-step abductive multi-target learning with diverse noisy samples (OSAMTL-DiNS) is proposed. • OSAMTL-DiNS is applied to tumour segmentation for breast cancer (TSfBC). • OSAMTL-DiNS enables various approaches for noisy labels achieving more rational results in TSfBC. • A predictive model pre-trained with OSAMTL-DiNS for TSfBC is released. Recent studies have demonstrated the effectiveness of the combination of machine learning and logical reasoning, including data-driven logical reasoning, knowledge driven machine learning and abductive learning, in inventing advanced technologies for different artificial intelligence applications. One-step abductive multi-target learning (OSAMTL), an approach inspired by abductive learning, via simply combining machine learning and logical reasoning in a one-step balanced multi-target learning way, has as well shown its effectiveness in handling complex noisy labels of a single noisy sample in medical histopathology whole slide image analysis (MHWSIA). However, OSAMTL is not suitable for the situation where diverse noisy samples (DiNS) are provided for a learning task. In this paper, giving definition of DiNS, we propose one-step abductive multi-target learning with DiNS (OSAMTL-DiNS) to expand the original OSAMTL to handle complex noisy labels of DiNS. Applying OSAMTL-DiNS to tumour segmentation for breast cancer in MHWSIA, we show that OSAMTL-DiNS is able to enable various state-of-the-art approaches for learning from noisy labels to achieve more rational predictions. We released a model pre-trained with OSAMTL-DiNS for tumour segmentation in HE-stained pre-treatment biopsy images in breast cancer, which has been successfully applied as a pre-processing tool to extract tumour-associated stroma compartment for predicting the pathological complete response to neoadjuvant chemotherapy in breast cancer. [ABSTRACT FROM AUTHOR]

Decentralized finance (DeFi) token scams are rising, leading to significant losses for investors and users. Hence, it is vital to identify malicious DeFi tokens accurately and early to protect investor funds. Current Machine Learning models that detect scam tokens rely entirely on DeFi token transactional behavior and have overlooked the public opinion that can be derived from news and social media posts, which can provide helpful information for identifying scam tokens. Further, existing frameworks to detect scam tokens fail to explain why certain tokens are identified as scams, hindering stakeholder confidence in such systems. This paper proposes a new transformer-based framework called "DeFiTrust" to identify malicious DeFi tokens using transaction events and social media posts. DeFiTrust has two components: event log processing and sentiment analysis. The former creates a feature vector representing temporal variations in the event logs of a token. The latter analyzes social media posts to determine public opinion about the token. A fully-connected neural network then processes both pieces of information to determine whether the token is legitimate or a scam. We evaluated the proposed model against existing Machine Learning-based models using real-world data. The results demonstrate that DeFiTrust outperforms existing approaches and detects scam tokens earlier. Explanations generated using Integrated Gradients reveal that DeFiTrust identifies scam tokens based on valid reasons, enhancing stakeholder trust in the proposed model. • DeFiTrust is a transformer-based framework that can identify malicious DeFi tokens. • Social media data has been incorporated to improve the detection ability. • DeFiTrust offers explanations for its decisions, improving stakeholder trust. • DeFiTrust can identify scam tokens before scams occur. [ABSTRACT FROM AUTHOR]

Vertical federated learning has become a novel approach to address the challenges of sharing training data and preserving data privacy in machine learning. However, there still exist privacy leakage risks in parameter transmission process while directly running vertical federated k -means clustering. To address these issues, this paper proposes a data reconstruction attack based on the distance matrices. The attack model does not affect the convergence speed and results of the model, nor does it disrupt the training process. The attacker only needs to receive, record, and analyze the transmitted distance matrices in the training process, and through steps, as establishment of new coordinate, possible location inference, target location extraction, coordinate alignment, and attribute filtering, can reconstruct the original training data set. Extensive experiments were conducted on three public datasets, the average relative Euclidean distances between the reconstructed dataset samples and the training dataset samples are 0.489%, 0.610%, and 0.887% for small-sample datasets, datasets with diverse data distributions, and high-dimensional datasets, respectively. Showing that the attack model is effective in the VF k -means clustering algorithm. • Vertical federated k-means clustering safeguards data privacy. • The distance matrix conveys pertinent information about the training dataset. • Simply replacing shared data cannot solve privacy leakages in VFk-means clustering. • It is possible to reconstruct the training dataset based on the distance matrix. • Reconstruction attacks do not disrupt the training process. [ABSTRACT FROM AUTHOR]

Many studies on predicting firm failure have focused on finding new features that improve the accuracy of the models. In this paper, genetic programming (GP) is used for this purpose. The main problem in GP is to specify a function that evaluates the fitness of the feature. Direct optimization of a machine learning (ML) model that uses a generated feature in most cases leads to high computational costs since evolving a population of N programs over G generations while evaluating each model using K -fold cross-validation requires N * G * K model learning cycles. Thus, many researchers use scores that measure the relationship of the generated features to the class label. However, our empirical analysis shows that most such scores correlate poorly with ML model performance. The novelty of our work is that we introduce several ways of combining different scores into a single measure of expected model performance. Experimental results on data from Hungarian firms (7167 observations, class imbalance 9.37) using five ML models (Logistic Regression, Random Forest, Gradient Boosting, Histogram Boosting, and AdaBoost) prove that the proposed way of setting the fitness function increases the ROC AUC of the listed models by 6.6%, 5.2%, 6.8%, 5.5% and 5.2% respectively. Moreover, by applying the found formula to the data from Czech firms (3872 observations, class imbalance of 74.92), which were not used for the feature search, we obtained increases in ROC AUC by 13.1%, 11.8%, 14.9%, 9.9%, and 8.2%, respectively. This indicates that the proposed method allows to find universal features, which opens the way to build effective models in case of insufficient data (small number of observations, extreme imbalance, etc.). [ABSTRACT FROM AUTHOR]

This paper aims to propose a new hybrid Machine Learning (ML) with Egret Swarm Optimization Algorithm (ESOA) and Weighted Relevance-based Combination Strategy (WERCS) methods for predicting the patch loading resistance (PLR) of both longitudinally unstiffened and stiffened plate girders. A dataset of both stiffened and unstiffened plate girders under patch loading are collected and used for training and testing to generate the proposed models. Firstly, the WERCS method is used to resample the training set to improve the performance of training data. After that, four ML models, including Gradient Boosted Tree (GTB), Extreme Gradient Boosting algorithm (XGB), Artificial neural network (ANN), and CATBoost regression (CAT) are employed. The hyperparameter of all above ML models are optimized by using ESOA algorithm to choose the best performance models. It was found that the WERCS influences considerably the performance of hybrid ML models, and the WERCS + ESOA + XGB model presents the best performance compared to others. The accuracy of the WERCS + ESOA + XGB model is validated by comparing its predictive results with the existing design codes and formulae. Additionally, the Local Interpretable Model-Agnostic Explanations (LIME) method is used to assess the importance and contributions of each input variable in the proposed model. This algorithm can explain the global and local effects of features with each range and create favourable foundations for practical designing. A Graphical User Interface (GUI) tool is developed to conveniently estimate both the patch loading resistance of longitudinally unstiffened and stiffened plate girders. A Python library has been developed and uploaded to PyPi for more convenient use without the need for accessing the source code. [ABSTRACT FROM AUTHOR]

Learning the discriminative features of different faces is an important task in face recognition. By extracting face features with neural networks, it becomes easy to measure the similarity of different face images, which makes face recognition possible. To enhance a neural network's face feature separability, incorporating an angular margin during training is common practice. The state-of-the-art loss functions CosFace and ArcFace apply fixed margins between the weights of classes to enhance the inter-class separation of face features. Since the distribution of samples in the training set is uneven, the similarities between different identities are unequal. Therefore, using an inappropriately fixed angular margin may lead to problems such as that the model has difficulty converging or that the face features are not sufficiently discriminative. It is more intuitive to use adaptive angular margins are angular adaptive, which can increase as the angles between classes increase. In this paper, we propose a new angular margin loss named X2-Softmax. X2-Softmax loss has adaptive angular margins, which increase as the angle between different classes increases. The angular adaptive margin ensures model flexibility and effectively improves the effect of face recognition. We trained a neural network with X2-Softmax loss on the MS1Mv3 dataset and tested it on several evaluation benchmarks to demonstrate the effectiveness and superiority of our loss function. [ABSTRACT FROM AUTHOR]

In traditional machine learning, there is often a discrepancy in data distribution between the source and target domains. Domain adaptation (DA) was proposed to learn the robust classifier for target domain by using knowledge from different source domains. Most DA methods focus on only the geometric structure of the data or statistical properties to reduce the differences between domains. The complementarity of these two aspects is ignored, which causes the problem of domain underadaptation to some degree. In this paper, we propose latent space search (LSS) approach for domain adaptation that consider both geometric and statistical properties. LSS consists of two parts: latent subspace learning and space search subspace learning. In the latent subspace, the geometric and statistical properties of the data are preserved by a low-rank coupled projection as well as joint distribution of the data. For the solution, an iterative feedback approach is used to obtain a robust subspace. Furthermore, to improve the discriminability of the subspace, the space search optimization algorithm is used to reconstruct the latent subspace, so that the source domain and the target domain can interleave well in the subspace. A comparative study illustrates that the proposed LSS has better performance than other state-of-art models reported in the literature. • Differences in data distribution exist in traditional machine learning. • Geometric and statistical properties are preserved in the latent subspace. • The discriminability of the subspace is raised in space search subspace. [ABSTRACT FROM AUTHOR]

Shoeprint reconstruction is essential in forensic science, but it is also challenging due to various inconsistencies in the patterns, textures, sizes, abrasions, etc. of shoeprints. The computational reconstruction of sharper and more complete shoeprints is conventionally conducted using handcrafted features, and it often requires human intervention. Prior studies using end-to-end machine learning approaches are limited in number and have not achieved a high level of performance. In this paper, we propose a model named ShoeRec, which employs variational autoencoder (VAE) as a component in a U-Net-like architecture to reconstruct missing regions and borders in shoeprint images. ShoeRec incorporates skip connections to preserve key patterns and employs VAE in the bottleneck to facilitate the reconstruction of desired shoeprints with the restoration of detail as perceived by humans. As a U-Net, the model skips the contextual information from the encoder to the decoder, and the compressed features in the latent space via VAE optimize the probabilistic distribution for reconstructing complete shoeprints. The reconstruction operation is automatically tuned according to the objective function, so as to reduce the structural correlation between the original and projected shoeprint and restore the absent information in the desired shoeprints. To the best of our knowledge, ShoeRec is the first deep learning infusion model that specializes in shoeprint reconstruction. The shoeprints reconstructed by ShoeRec have a close match with the originals in terms of both structures and patterns, and ShoeRec outperforms state-of-the-art generative models in human evaluation. • An end-to-end deep learning model focuses on shoeprint reconstruction for forensics. • Reconstruction of existing patterns together with the generation of missing regions. • Missing regions reconstruction is specifically designed based on the loss function. • The proposed model performs well for noisy, blurry, and partial input images. • Model outperforms state-of-the-art generative models in terms of human evaluation. [ABSTRACT FROM AUTHOR]

Portfolio management (PM) is a central subject of finance that presents many unresolved challenges. Although solutions have been proposed since the 1950's, recent advances in machine learning are revolutionizing investments. Whereas supervised paradigms such as regression and classification are established in this direction, emerging research suggests that the learning-to-rank (LtR) paradigm can be more effective. Yet, these works are not many and overlook some important aspects. These aspects include the risk inherent in stocks, feature representation, portfolio cardinality and allocation. Addressing these gaps, this paper presents Stochastic-Aware Bootstrap Ensemble Ranking (SABER), an LtR method that directly minimizes the uncertainty and improves ranking performance. In addition, we propose Merged Bootstrap Selection for dynamic portfolio cardinality optimization. Trained on macroeconomic, fundamental and market indicators, the proposed methods outperform a state-of-the-art LtR approach, reaching annual returns and volatility of 83.48% and 29.89% across 6 years. The former also considers renowned weight allocation methods and achieves Sharpe and Sortino ratios of 2.79 and 5.28 respectively. Finally, the computational complexity of the proposed method is proven comparable to that of current methods. • Comprehensive approach to stock ranking, portfolio cardinality and weight allocation. • Risk-aware stock ranking method based on the learning-to-rank (LtR) paradigm. • Effective integration of market, fundamental and macroeconomic features. • Experiments with five renowned portfolio allocation methods. • The proposed strategy achieved Sharpe and Sortino ratios of 2.79 and 5.28. [ABSTRACT FROM AUTHOR]

As an increasing number of investors share their opinions on social networks, a critical challenge is to provide advice and assistance in making well-informed investment decisions by considering enormous online sentiments. A typical way is to aggregate investors' sentiments into an investing signal. Most previous studies of sentiment aggregation rely on human-defined rules summarized from experience, which are widely accepted for simplicity and interpretability. However, the relations between investor sentiments and wise investment decisions generated by integrating sentiments could be complex and counter-intuitive, which may not be fully described by human-defined rules and require further investigation. This paper proposes a novel machine learning-based method called FollowAKOInvestor to integrate investors' sentiments more effectively for highly profitable stock recommendations. It divides investors into groups by clustering algorithm or their expertise levels, then extracts sentiment features from each group to train a machine learning model. The outputs are utilized to score stocks and provide investment recommendations. Based on an extensive dataset spanning from January 2017 to December 2019, sourced from StockTwits, we simulate a one-year recommendation procedure to compare and evaluate the performance of the proposed method and baselines. The experimental results support the superiority of machine learning over human-defined rules in making the most of sentiments from various investors for stock recommendations. • A novel machine learning-based method is proposed to aggregate investor sentiments. • Sentiments from various investors matter for the stock recommendation. • Applying machine learning to follow various investors boosts performance. • Comparative results on a real-world dataset corroborate the efficacy of our method. • It provides implications for investment practice and crowd intelligence generation. [ABSTRACT FROM AUTHOR]

• The paper explores AI decision-making to enhance transparency of CIN diagnosis using colposcopy images. • The study presents interpretable insights for clinical validation using different XAI methods. • Saliency method showed best result, using gradients to reveal likely malignancy traits. • Expert Clinic Validation of results conducted, along with statistical analyses. Although artificial intelligence techniques have performed well in analysing medical images, it is essential to consider the degree of understanding or interpretation in evaluations. This is especially important when diagnostic support is involved, as the trustworthiness of machine learning predictions needs to be clarified and evaluated. For deep learning algorithms to be translated into clinical practice, they must be made less opaque. To examine the classification of cervical dysplasia-diagnosed patients, we will explore explainable artificial intelligence (XAI) approaches, by implementing seven different interpretation techniques (Saliency, XRAI, Integrated Gradients, Smooth Gradients, Smooth Integrated Gradients, Grad-CAM, Smoothgrad-CAM). These XAI techniques shall be explored in order to assess the performance of four Convolutional Neural Network models (AlexNet, VGG16, MobileNet, and InceptionNet). Our results show that saliency qualities most closely match expert annotations, and there is a moderate to strong association between a model's sensitivity and the agreement between humans and computers. We also discovered a relevant relationship between expert insights and computational learning. This highlights the potential role of human expertise in influencing effective computer learning. [ABSTRACT FROM AUTHOR]

Surface roughness is a critical parameter for quantifying the surface quality of a workpiece. Accurate surface roughness prediction can significantly influence the overall quality of the final components by reducing costs and enhancing productivity. Although various prediction methods have been explored in previous research, less attention has been given to explainable prediction methods for surface roughness. This paper proposes a surface roughness prediction model based on the Differential Evolution (DE) algorithm with ensemble learning. It aims to elucidate the intrinsic correlation between processing parameters and surface roughness values in Multi-Jet Polishing (MJP) using explainable analysis methods. The proposed Ensemble Regression with Differential Evolution (ERDE) algorithm comprises two main modules: data processing and analytics, and ensemble model prediction. These modules investigate the surface roughness of MJP from the data and model levels, respectively. To validate the effectiveness of ERDE, MJP experiments were conducted on three-dimensional printed components of 316L stainless steel. The experimental data indicated that ERDE outperforms other existing algorithms, reducing the mean absolute percentage error by approximately 42 %. [ABSTRACT FROM AUTHOR]

Recently, researchers have been extensively exploring the immense potential of Q-Star. However, the available resources lack comprehensive information on this topic. Despite this, a Q-table, an uncomplicated lookup table containing actions and states, is often seen as a mere data structure for tracking. This overlooks the vast amount of knowledge that can be derived from it through visualization. The Q-learning algorithm utilizes this table to update values and determine the highest anticipated rewards for actions in each state. However, instead of relying solely on complex reward functions for algorithm development, leveraging the existing knowledge within the Q-table would be highly beneficial. Incorporating this valuable information into the algorithmic framework can minimize the need to develop intricate reward functions. This paper proposes an expected-mean gamma-incremental Q approach to tackle the challenges of convergence speed in an uninformed search reinforcement learning (RL) algorithm and the issue of path optimality in path planning problems. The gamma-incremental RL method revolves around adjusting the weight of the future value by considering the level of exploration. It enables the robot to receive preference feedback, either near-term reward or long-term reward, based on the frequency of the visited state. Meanwhile, the expected-mean technique uses the information of the robot's turning actions to update the Q-target. By consistently incorporating valuable insights from the Q-table, the algorithm can gradually enhance its understanding of the available information, resulting in more efficient decision-making. The experiment results indicate that the proposed algorithm accelerates the convergence rate, outperforming the baseline Q-learning by up to 2 times. It addresses the challenge of robot path planning by prioritizing promising solutions, resulting in near-optimal outcomes with higher total rewards and enhanced learning stability. [ABSTRACT FROM AUTHOR]

Feature selection has a prominent role in high-dimensional datasets to increase classification accuracy, decrease the learning algorithm computational time, and present the most informative features to decision-makers. This paper proposes a two-stage hybrid feature selection for high-dimensional medical datasets: Maximum Pattern Recognition - Multi-objective Discrete Evolution Strategy (MPR-MDES). MPR is a rapid filter ranker that significantly outperforms existing frequency-based rankers in recognizing non-linear patterns, effectively eliminating a majority of non-informative features. Then, the wrapper Multi-objective Discrete Evolution Strategy (MDES) uses the remaining features and obtains sets of solutions which are automatically presented to decision-makers. The experiments conducted on large medical datasets demonstrate that MPR-MDES achieves considerable improvements compared to state-of-the-art methods, in terms of both classification accuracy and dimensionality reduction. In this sense, the proposal successfully performs when presenting informative feature sets to decision-makers. The implementation is available on https://github.com/KhaosResearch/MPR-MDES. [Display omitted] • Maximum Pattern Recognition (MPR) is proposed as a frequency-based filter ranker. • Multi-objective Discrete Evolution Strategy (MDES) is proposed as a wrapper method. • The proposed MPR-MDES selects the best feature sets in the Pareto front. [ABSTRACT FROM AUTHOR]

Intelligent Decision Support Systems (DSS) have gained considerable interest within the field of construction engineering through the utilization of Artificial Intelligence (AI) and Machine Learning (ML) methodologies. These systems possess the capacity to enhance decision-making processes and optimize the results of projects. Nevertheless, it is imperative to conduct a thorough and inclusive examination that incorporates and evaluates the various methodologies employed within this particular discipline, thereby bridging the existing void in research. The current state of research in the field of AI and ML in construction engineering is characterized by a dearth of comprehensive reviews that encompass the wide array of methodologies utilized. Although certain studies have delved into specific facets of this subject, there remains a novelty and research gap in terms of a holistic examination of the topic. The objective of this review is to conduct a comprehensive analysis of AI and ML methodologies in the field of construction engineering, with a specific emphasis on their practical applications, advantages, and constraints. The study also ascertains potential areas for future research and improvement in the domain of Intelligent DSS. This study employed a rigorous approach that involved an extensive review of scholarly literature and empirical investigations from 2001 to June 2023. The findings of the analysis indicate a noticeable increase in the number of research papers focusing on Intelligent DSS within the field of construction engineering during the last twenty years. The field has witnessed the emergence of key research areas, namely the utilization of Internet of Things (IoT), explainable AI, supply chain optimization, sustainable construction practices, and human–machine collaboration. The results of this study hold significant implications for the construction industry. The incorporation of AI and ML techniques within Intelligent DSS can effectively tackle various issues such as cost overruns, project delays, and safety concerns. [ABSTRACT FROM AUTHOR]

This paper proposes an intelligent risk management system in stock markets based on indications of social media platforms. Based on a brief survey, we found that the literature focuses on identifying, assessing and optimizing risks in stock markets using classical data sources as well as utilizing mathematical, statistical and machine learning techniques. So, we suggest a trendy data source that is social media and employing natural language processing as an advanced technique to identify and estimate risks in stock markets. This NLP model can help the investors of stock markets to be aware of the potential risks, its types and magnitudes which can help them to take the best decisions. In addition , we describe this model and some important expressions for risk analysis, risk recognition and risk estimation are driven. Besides, a practical experiment is provided to identify and evaluate the potential risks in NASDAQ stock market. Moreover, several performance measures are used to validate the efficiency of the proposed model such as the mean accuracy for the processes of risk analysis, risk identification and risk assessment is 73.33 %. [ABSTRACT FROM AUTHOR]

Correctly predicting up trends, sideways trends and down trends for stock returns is vitally important in financial market. Group penalized trinomial logit dynamic models can not only provide three-class label information and three-class probability estimations, but also enhance three-class prediction performance by shrinking group coefficients to bypass multi-collinearity and over-fitting. In this paper we propose G-LASSO/G-SCAD/G-MCP penalized trinomial logit dynamic models with 24 technical indicators to predict up trends, sideways trends and down trends for stock returns, develop group coordinate descent algorithm (GCD) to complete group selection and group estimation simultaneously, establish the relative optimal Bayes classifier to identify class labels, introduce three-class confusion matrix and hypervolume under the ROC manifold (HUM) to assess the three-class prediction performance to 15 methods. Experiment results show that G-LASSO/G-SCAD/G-MCP penalized trinomial logit dynamic models predict better than LASSO/SCAD/MCP penalized trinomial logit dynamic models, 6 deep learning models and 3 machine learning models in terms of Accuracy and HUM. In particular, the highest prediction Accuracy (the average Accuracy) from G-LASSO penalized method outperforms 3-Layer Long Short-Term Memory (LSTM) and Random Forest (RF) for 4.76% and 10.29% (8.16% and 11.78%), respectively. Moreover, compared with LASSO/SCAD/MCP penalized methods, G-LASSO/G-SCAD/G-MCP penalized methods enhance 5.88% Accuracy and only require 27.29% Predicted Time. The proposed G-LASSO/G-SCAD/G-MCP penalized trinomial logit dynamic models can not only be extended as more general three-class prediction method by replacing 24 technical indicators by some factors influencing three-class response variable, but also be directly extended as more efficient multi-category prediction method by replacing three-class response variable by multi-category response variable. • Propose the new group penalized methods to predict stock return movement direction. • Develop a new GCD algorithm for G-LASSO/G-SCAD/G-MCP penalized trinomial logit model. • Obtain the new rapidly convergent G-LASSO/G-SCAD/G-MCP estimator. • Introduce a few measures to assess three-class prediction performance to 15 methods. • G-LASSO/G-SCAD/G-MCP penalized methods predict better than the other 12 methods. [ABSTRACT FROM AUTHOR]

Accurately forecasting demand is a potential competitive advantage, especially when dealing with perishable products. The multi-billion dollar horticultural industry is highly affected by perishability, but has received limited attention in forecasting research. In this paper, we analyze the applicability of general compared to dataset-specific predictors, as well as the influence of external information and online model update schemes. We employ a heterogeneous set of horticultural data, three classical, and twelve machine learning-based forecasting approaches. Our results show a superiority of multivariate machine learning methods, in particular the ensemble learner XGBoost. These advantages highlight the importance of external factors, with the feature set containing statistical, calendrical, and weather-related features leading to the most robust performance. We further observe that a general model is unable to capture the heterogeneity of the data and is outperformed by dataset-specific predictors. Moreover, frequent model updates have a negligible impact on forecasting quality, allowing long-term forecasting without significant performance degradation. • A comprehensive study on time series forecasting for horticultural sales. • Machine learning outperforms classical forecasting in horticultural sales prediction. • Statistical and calendar features are beneficial for horticultural sales prediction. • Detailed analysis of the impact of model refitting schemes on forecasting quality. • Company-specific forecasting models are superior to cross-company predictors. [ABSTRACT FROM AUTHOR]

The volume of e-commerce continues to increase year after year. Buying goods on the internet is easy and practical, and took a huge boost during the lockdowns of the Covid crisis. However, this is also an open window for fraudsters and the corresponding financial loss costs billions of dollars. In this paper, we study e-commerce credit card fraud detection, in collaboration with our industrial partner, Worldline. Transactional companies are more and more dependent on machine learning models such as deep learning anomaly detection models, as part of real-world fraud detection systems (FDS). We focus on continual learning to find the best model with respect to two objectives: to maximize the accuracy and to minimize the catastrophic forgetting phenomenon. For the latter, we proposed an evaluation procedure to quantify the forgetting in data streams with delayed feedback: the plasticity/stability visualization matrix. We also investigated six strategies and 13 methods on a real-size case study including five months of e-commerce credit card transactions. Finally, we discuss how the trade-off between plasticity and stability is set, in practice, in the case of FDS. • Fraud detection models must be updated continually to handle new fraud strategies. • They must balance plasticity (learn new patterns) and stability (remember old ones). • We show how to quantify both and discuss the trade-off for fraud detection. • We provide an extensive comparison of six strategies and 13 different models. • We present a real case study based on more than 50 million e-commerce transactions. [ABSTRACT FROM AUTHOR]

Visual features encompass visual information extracted from images or videos, serving as supplementary input to enhance the efficacy of neural machine translation (NMT) systems. By seamlessly integrating these visual features into the translation process, NMT models can tap into the expansive visual context, thereby making more precise predictions during translation. Nonetheless, limited efforts have been directed towards exploring the potential of learning target-side future context using visual features to augment NMT performance. To bridge this gap, this paper introduces the Vision-guided Target-side Future Context (VisTFC) learning framework for NMT. Our core objective is to refine translation quality by effectively harnessing and incorporating target-side future contextual insights derived from the visual modality. The VisTFC framework consists of three pivotal components. Firstly, a graph-based multimodal encoder–decoder is established, constructing a bipartite vision-source/target graph. This enables the acquisition of vision-fused textual representations, synthesizing both linguistic and visual attributes to enhance translation accuracy. Secondly, a target-side future context predictor, incorporating a dynamic routing mechanism, infers future context by synergizing visual information. This empowers the model to anticipate and assimilate contextual cues, ensuring more coherent and contextually coherent translations. Thirdly, a sigmoid update gate is introduced, facilitating controlled integration of predicted future context into the decoding process, enabling the decoder to flexibly adapt and utilize the inferred context. Moreover, the VisTFC framework is fortified with additional loss functions that enforce source/target-vision consistency, reinforcing its robustness and effectiveness. Our results demonstrate that our model can achieve substantial improvements, reaching up to 1.0/0.9/1.1 BLEU points beyond the performance of the most robust baselines. This superior performance is evident on both the Test2016 and Test2017 datasets, as well as on the MSCOCO test sets for English-French translation tasks. Furthermore, our experiments underscore the efficacy of VisTFC, even in scenarios with limited resources, such as the English-Hausa language pair. Here, the model achieves notable enhancements, with improvements of up to 1.8/1.9 BLEU points compared to the strongest text-only NMT models on E-Test and C-Test test sets. These findings provide strong evidence supporting our model's ability to effectively harness target-side future contextual information from the visual modality, resulting in substantial enhancements in machine translation quality. Further analysis and visualization suggest that VisTFC is capable of learning target-side future context from visual signals for better translation. We also have open-sourced our work in Github. 1 1 https://github.com/nlpdl/VsiTFC. • Exploring how to incorporate visual features into the translation process. • Adopting a dynamic routing mechanism to infer target-side future context. • Integrating a multimodal graph into encoder–decoder. [ABSTRACT FROM AUTHOR]

• We predict the number of shares of fake news shares based on its content. • The propagation of fake and real news follows different patterns. • Fake news that conveys strongly negative emotions are more likely to be shared. • We use machine learning to detect early fake news that are potentially viral. Fake news about brands and companies can cause reputation and financial damage, but there exist preventive strategies that could be implemented if we were able to early predict the virality of fake news pieces. In this paper, we consider the early prediction of fake news virality using textual information (body and headlines). We show that the impact of content features on virality is different for true and fake news, and so the propagation of true and fake news is. Furthermore, we use machine learning nonparametric models to classify fake news according to its propagation level. We also discuss defensive response strategies. [ABSTRACT FROM AUTHOR]

Imbalanced class distributions in machine learning, where the minority class is often under-represented, pose a substantial challenge. Synthetic Minority Over-sampling Technique (SMOTE) has been widely employed to address this issue by generating synthetic minority samples through interpolation. Despite its popularity, SMOTE exhibits certain drawbacks caused by the implementation of random interpolation samples. In this paper, we introduce a new data level technique for oversampling, called Fuzzy C-Means Center-SMOTE (FCM-CSMOTE), which generates synthetic samples in each cluster using its center considered as the memory of the main data components. We demonstrate that the proposed selective strategy has a very low probability to generate noise. The experimental results demonstrate that the proposed method performs better than the state-of-the-art approaches on 21 real unbalanced data sets (regular and large size data set) in terms of several metrics, including Geometric Mean (GM), F-Measure (FM), Area Under the Curve (AUC), and Accuracy. [ABSTRACT FROM AUTHOR]

This paper presents a novel approach that merges a lightweight parallel depth-wise separable convolutional neural network (LPDCNN) with a ridge regression extreme learning machine (Ridge-ELM) for precise classification of three lung cancer types alongside normal lung tissue (adenocarcinoma, large cell carcinoma, normal, and squamous cell carcinoma) using CT images. The proposed methodology combines contrast-limited adaptive histogram equalization (CLAHE) and Gaussian blur to enhance image quality, reduce noise, and improve visual clarity. The LPDCNN extracts discriminant features while minimizing computational complexity (0.53 million parameters and 9 layers). The Ridge-ELM model was developed to enhance classification performance, replacing the traditional pseudoinverse in the ELM approach. Through comprehensive evaluation against state-of-the-art models, the framework achieves remarkable average recall and accuracy values of 98.25 ± 1.031 % and 98.40 ± 0.822 %, respectively, through rigorous five-fold cross-validation for four-class classifications. In binary classifications, outstanding results are obtained with recall and accuracy values of 99.70 ± 0.671 % and 99.70 ± 0.447 %%, respectively. Notably, the framework exhibits exceptional efficiency, with a testing time of only 0.003 s. Additionally, integrating the SHAP (Shapley Additive Explanations) in the proposed framework enhances Explain-ability, providing insights into decision-making and boosting confidence in real-world lung cancer diagnoses. [ABSTRACT FROM AUTHOR]

Nowadays, online community platforms provide firms with an important source of information for conducting dynamic marketing research. High-technology companies, in particular, rely heavily on lead users for the development of very novel products or easily adjustable services. In this paper, we present a three-phase methodology that integrates a machine-learning-based algorithm with a sophisticated clustering technique. The purpose of this methodology is to systematically identify lead users and their needs from a complex online community network. We also aim to identify important features, perceptions, and preferences for different groups of lead users. To validate the effectiveness of our approach, we conduct a real-world case study. • Lead users in online communities are valuable for new product development. • A novel three-phase framework is proposed to identify lead users and their needs. • Lead users' need trends are extracted by a sophisticated clustering technique. • Comparative and industrial studies validate the superiority of the method. [ABSTRACT FROM AUTHOR]

Active learning is focused on minimizing the effort required to obtain labeled data by iteratively choosing fresh data samples for training a machine learning model. One of the primary challenges in active learning involves the selection of the most informative instances for labeling by an annotation oracle at each iteration. A viable approach is to develop an active learning strategy that aligns with the performance of a meta-learning model. This strategy evaluates the quality of previously selected instances and subsequently trains a machine learning model to predict the quality of instances to be labeled in the current iteration. This paper introduces a novel approach to learning for active learning, wherein instances are chosen for labeling based on their potential to induce the most substantial change in the current classifier. We explore various strategies for assessing the significance of an instance, taking into account variations in the learning gradient of the classification model. Our approach can be applied to any classifier that can be trained using gradient descent optimization. Here, we present a formulation that leverages a deep neural network model, which has not been extensively explored in existing learning-to-active-learn methodologies. Through experimental validation, our approach demonstrates promising results, especially in scenarios where there are limited initially labeled instances, particularly when the number of labeled instances per class is extremely limited. • A new meta-active learning strategy for deep neural network models. • The instance selection problem is modeled as a regression problem. • Four strategies of instance importance scoring by considering gradient variations. • Experiments have shown good results when few initially labeled objects are available. • A dimensionality reduction step for the instances improves the performances. [ABSTRACT FROM AUTHOR]