Abstract The proposed framework aims to address the challenge of high-dimensional, noisy, irrelevant, and redundant features in clinical datasets by using a two-step feature selection approach. The first step involves a wrapper-based method for calculating the Permutation Based Feature Importance (PBFI) score using the Support Vector Machine (SVM) classifier. The second step employs three wrapper-based bio-inspired optimization algorithms, namely Harris Hawk Optimization (HHO), Binary Grasshopper Optimization (BGOA), and Whale Optimization (WOA), with the weighted F1-Score, measured by the SVM classifier, as the fitness function. The selected features from both steps are then combined using a union operation and used to train four classifiers: SVM, K-Nearest Neighbor (K-NN), Linear Discriminant Analysis (LDA), and Naive Bayes (NB). The proposed framework is also evaluated using a non-parametric hypothesis testing method, the Kruskal-Wallis Test, on eight medical datasets from the Machine Learning Repository (MLR) maintained by the University of California, Irvine (UCI). The proposed approach is compared with feature selection using the Particle Swarm Optimization algorithm (PSOA) and Lion's Algorithm, and is shown to perform better.
Most existing research on cyberbullying detection reduces the task to comment-level hate speech classification, overlooking its collective, event-driven, and dynamic nature. To address this limitation, we propose CDMPL, a multi-feature prompt learning framework for Chinese cyberbullying incident detection. CDMPL integrates textual features (semantic summaries generated by large language models) with temporal features (temporal behavioral patterns captured by a Dynamic Attention Truncation mechanism and an Autoformer-based time series encoder). These fused features are reformulated as a masked language modeling task, enhancing interpretability and robustness. We evaluate CDMPL on the only publicly available Chinese incident-level dataset, consisting of 86 cyberbullying and 109 non-cyberbullying events. Under a 15-shot setting, CDMPL achieves 84.79% accuracy and 84.05% F1-score, exceeding state-of-the-art large language models such as Doubao (71.91% F1) and ChatGPT-4o (70.06% F1) by more than 12%. Ablation studies further demonstrate the necessity of jointly modeling semantic and temporal features. To our knowledge, this is the first systematic study of Chinese incident-level cyberbullying detection, providing methodological and empirical advances for early detection and governance.
Abstract Underwater object detection (UOD) is a major process in marine exploration, ecological monitoring, autonomous underwater vehicles (AUVs) and ocean observation systems. However, underwater environments are affected by low contrast, color attenuation, light and blurred object boundaries, which degrade the performance of conventional UOD models. Practical underwater applications need lightweight architectures capable of deployment on resource limited platforms. For addressing these problems, this work suggests UEAFM-YOLO, a lightweight UOD model developed on the YOLOv11n network. First, an Underwater Edge-Aware Feature Module (UEAFM) module is introduced to enhance edge-aware representations and improve feature extraction under visually degraded underwater conditions. Second, a Cross-Stage Partial with Lightweight Star Fusion (C3k2LSF) is presented for reducing computational complexity and preserving feature representation capability. Third, an Underwater Multi-Scale Semantic Feature Pyramid Network (UMSFPN) is presented for strengthening semantic consistency and cross-scale feature interaction, which makes robust UOD. The proposed model integrates low-level structural features with high-level semantic features for improving object localization and classification performance. Experimental analysis demonstrates that UEAFM-YOLO achieves better mAP@0.5 values of 91.0% on DUO and 90.5% on UTDAC2020 datasets. Thus, the proposed model is highly suitable for marine surveillance, underwater robotics, and unmanned survey systems.
Interval-valued neutrosophic preference relations (IVNPRs), as an extension of interval-valued neutrosophic sets in preference modeling, can represent DMs’ hesitation and uncertainty through interval-form truth, indeterminacy, and falsity membership degrees. However, existing IVNPR-based group decision-making (GDM) models often separate consistency rectification from consensus reaching or cause information loss during ranking. This paper proposes an integrated GDM framework for consistency, consensus, and ranking based on IVNPRs. First, an additive consistency index (ACI) is defined, and a parametric linear programming model is developed to minimize modifications to original preferences. Second, a consensus optimization model is constructed to jointly adjust preferences and determine DM weights while reflecting expert reliability. Third, a likelihood comparison-based ranking method is designed by integrating exponential distance, TOPSIS, and interval inclusion information to reduce information loss. Finally, sensitivity analysis, comparative analysis, ablation analysis, and computational efficiency analysis are conducted to verify the model’s stability, applicability, and scalability. The proposed framework provides a robust tool for intelligent decision-making by preserving original expertise while resolving logical inconsistencies.
Abstract Unmanned aerial vehicle (UAV) systems, along with Artificial intelligence (AI) algorithms, are being increasingly advocated for management and decision-making processes in mining and construction industries. This paper presents a concise literature survey of recent works in this regard while drawing some useful insights. We mainly focus on areas like terrain modeling, volumetric analysis, autonomous inspection, hazard detection and integration with digital infrastructure. Our survey highlights that we can significantly improve volumetric accuracies and defect detection rates. Moreover, real-time processing can be attained that is very useful for operational deployment. The paper also presents some of the key challenges and emerging trends associated with AI-enabled drone systems in mining and construction.
Carotid ultrasound image classification plays a vital role in carotid plaque diagnosis and stroke risk prediction. However, mixed-echoic plaques are difficult to classify because of low contrast between plaques and surrounding tissues and insufficient global contextual modeling. To address these issues, this paper proposes a Contrastive Enhancement and Context Guided Fusion Network (CEGF-Net) for carotid plaque classification. The proposed network contains a Contrast-Enhanced Feature Fusion (CEFF) module and a Context Guided Fusion (CGF) module. CEFF enhances the contrast between plaque foreground and surrounding tissue background, while CGF adaptively integrates local details and global contextual cues to improve the representation of ambiguous plaque regions. Experiments on 1270 carotid plaque ultrasound images collected from a collaborating hospital show that CEGF-Net improves classification accuracy by approximately 1.9% compared with the baseline network. In particular, the accuracy for mixed-echoic plaque classification increases by around 9%, demonstrating the effectiveness of the proposed approach.
Abstract One of the fastest globally proliferating malignancies is skin melanoma cancer. Thus, by using dermoscopy images, researchers hope to develop a reliable technique for identifying skin lesions as malignant or benign, which they can use in clinical practice further. This paper proposes a novel way of identifying skin melanoma by fusion of hand-crafted features together with deep learning features followed by feature selection step. A custom dataset stated on Kaggle as the Melanoma Skin Cancer Dataset comprising of 10605 dermoscopic images of malignant as well as benign lesions is used, on which various preprocessing steps are performed followed by feature extraction and fusion. The hand-crafted GLCM, GLRLM, SFTA features are extracted and combined with deep learning features from DenseNet201. The features are then selected from fused feature set using bio-inspired based Whale Optimization Algorithm (WOA). Five cross-validation protocols (k = 2, 3, 4, 5, 10) are carried out on the dataset. The Kernel-SVM classifier is used for the classifying the skin lesions as benign or malignant. Under 10-fold cross-validation, the proposed model achieves the mean accuracy of 99.28%, sensitivity of 99.28% and specificity of 99.27% outperforming individual feature models and PSO-based feature selection. A statistical significance testing using paired t-test confirms that the improvements over comparative techniques are significant. Further validation on the external HAM10000 dataset demonstrates strong generalization capability achieving 96.52% accuracy. The proposed fused features WOA-SVM framework provides an efficient solution for melanoma classification.
Deepfake technology employs state-of-the-art deep learning to create hyper-realistic synthetic media, which has critical implications for digital authenticity and information security. The spread of such faked content has dire social consequences, facilitating misinformation campaigns, identity theft, and loss of trust in digital media. Today's detection systems experience limitations, such as single-model architectures tending to be unable to generalize the variety of artifact patterns in varying deepfake generation techniques, and most methods are not robust to adversarial attacks. To address these challenges, this study introduces an ensemble-based approach, namely the X-Iv2 Ensemble approach, merging Inception ResNet v2 and Xception Net based on their complementary architectures to enhance feature extraction and classification. The approach utilizes a region-based preprocessing technique, dividing input faces into Upperface, Lowerface, and Eyes regions using facial landmarks on four varied datasets such as DFDC, Celeb-DF, FF++ and SDFVD to ensure robust generalization. Notably, the inclusion of varied races, ages, and genders in the dataset ensures equitable performance across demographics. Apart from these improvements, most deep learning detection frameworks are still black boxes with limited interpretability and practical applications. This approach integrates Explainable AI (XAI) using Integrated Gradients to interpret decision-making processes, while adversarial testing through input perturbations comprehensively assesses the robustness of the model. Extensive assessment metrics such as accuracy (97%), sensitivity (97%), and specificity (96%) illustrate the system's better performance in balancing deepfake detection effectiveness and generalization across datasets.
Large Language Models (LLMs) have demonstrated remarkable proficiency in natural language processing applications, encompassing question answering, text generation, and reasoning capabilities. However, their metacognitive abilities, which involve self-assessment, uncertainty awareness, and cognitive control, remain insufficiently explored. This investigation examines over 97 publications released between 2021 and 2025. These studies encompass prompt-based methodologies, fine-tuning techniques, retrieval-augmented generation, and agentic AI frameworks that implement metacognition within lifelong learning models (LLMs). Metacognitive interventions produce measurable, task-dependent performance gains. Reported improvements range from approximately 3% to over 20% across evaluated benchmarks. These gains are observed in medical reasoning, multi-hop question answering, and natural language comprehension tasks. These enhancements correlate with demonstrable improvements in confidence calibration, error detection, and the precision of response revisions. Furthermore, this work conducts a comprehensive examination of more than ten families of large language models and agentic frameworks, systematically identifying enduring challenges, including overconfidence, susceptibility to hallucinations, limited error awareness, and unstable self-reflection processes. This research furnishes an analytical basis for elucidating the conditions under which metacognitive mechanisms either enhance or diminish LLM reliability. Moreover, it delineates prospective research avenues aimed at developing scalable, trustworthy, and human-centered artificial intelligence systems. This foundation is established through the synthesis of evaluation protocols, benchmarking methodologies, and comparative evidence.
The rapid growth of artificial intelligence and machine learning applications in project management is transforming the prediction of key performance indicators (KPIs). This has led to increased interest from both practitioners and the academic community, as reflected in the growing number of publications, creating a need for a comprehensive review. This paper presents a systematic scoping review of 688 peer-reviewed publications spanning three decades, analyzing the application of machine learning methods for predicting project KPIs across industries. The study aims to provide a comprehensive overview of current knowledge, offering insights and future research directions to advance the field and encourage greater innovation and efficiency in project management. The review focuses on identifying key patterns and trends in the data and exploring the potential of machine learning to enhance project outcomes. It also addresses the challenges and limitations of adopting these techniques. It reveals that machine learning applications are predominantly focused on cost and schedule performance prediction, with neural network-based approaches emerging as the most widely used techniques across industries. The findings indicate consistent improvements in predictive performance in complex project environments, while highlighting persistent challenges in data quality, model interpretability, and integration with existing project management systems.
Contrastive learning has become a powerful paradigm for unsupervised representation learning. However, its effectiveness largely depends on carefully designed data augmentation strategies to generate meaningful positive and negative pairs. Additionally, unsupervised clustering algorithms are typically sensitive to initialization and prone to converging to suboptimal local minima, resulting in unstable performance. To overcome these challenges, we propose HAPL, a unified end-to-end framework for short text clustering that integrates Hybrid data Augmentation with Pseudo-Label supervision. HAPL combines explicit and implicit data augmentation techniques in a synergistic strategy. It also incorporates an adaptive optimal transport mechanism for pseudo-label generation. This design provides principled supervision that stabilizes the optimization process and adapts to varied cluster distributions, thereby enhancing the model’s discriminative power. Furthermore, prototype learning is employed to reinforce the coherence of representations in the embedding space. Extensive experiments on eight benchmark datasets show that HAPL achieves state-of-the-art performance across various evaluation metrics. Comprehensive ablation experiments validate the contribution of each component to the overall effectiveness of the framework.
Incorporating classic literature as a part of language learning is not only engaging but also interesting for English as a Foreign Language (EFL) learners. It immerses them in the narrative, sparking their curiosity and motivation to keep reading. However, the rich vocabulary found in these classics may pose a significant challenge and surpass the proficiency levels of learners. This study proposes a novel mechanism called Adaptive Language Learning by Leveraging BERT and Semantic Technologies (ALBS) to address this issue. The proposed ALBS aims to enhance customized lexical proficiency for EFL learners. It replaces challenging vocabulary with words better aligned with learners’ lexical capabilities while preserving the original semantic meaning and syntactic fluency. It facilitates the recommendation of classic literature to EFL learners at varying vocabulary proficiency levels, enabling tailored learning experiences and suitable lexical competency, which enhances learner motivation. The proposed ALBS is structured into three distinct phases to achieve this goal. In the first phase, an L-BERT model is trained using Natural Language Processing (NLP) techniques, aiming to identify the level of each word from the input sentence. Then the distributions of the vocabulary proficiency levels of both classic literature and the learner are mapped out in percentages through L-BERT. Finally, in the word replacement phase, the three criteria of fluency, semantics, and the Common European Framework of Reference for Languages (CEFR) word level are integrated to select the most suitable word from the list as a replacement for the target word. The results indicate that the proposed ALBS outperforms the existing mechanisms in terms of precision, recall, and F1-score.
Automatic Identification System (AIS) data plays a critical role in maritime analytics; however, AIS trajectories frequently suffer from noise, inconsistencies, and missing values that degrade their analytical reliability. To address these challenges, this study proposes the Deep Quality Vessel Trajectory Inspector (DQTI), a hybrid variational framework for AIS data quality enhancement and anomaly detection. The proposed model integrates a Variational Recurrent Neural Network (VRNN) with a learnable fusion of GRU and LSTM units, enabling effective modeling of both short-term motion continuity and long-range temporal dependencies in vessel trajectories. A four-hot encoding scheme is adopted to represent longitude, latitude, speed over ground, and course over ground, providing a structured and noise-tolerant representation of multivariate maritime signals. Anomaly detection is performed through a reconstruction-based mechanism that identifies inconsistent AIS messages by measuring deviations between observed and reconstructed trajectories. The experimental evaluation is conducted on more than 141,000 AIS records collected from 800 vessels in the Red Sea. Results demonstrate improved reconstruction accuracy, reduced KL divergence, and more reliable anomaly discrimination compared to baseline models. These findings highlight the effectiveness of data representations and hybrid variational recurrent architectures for enhancing AIS data quality in complex and noisy sequential datasets.
Traffic accidents involving heavy vehicles remain a critical safety challenge, with human factors being the primary contributors to fatality risks. While machine learning has been widely used for accident prediction, the lack of model interpretability often hinders its application in real-world policy-making. This study proposes a robust intelligent framework to classify truck driver fatalities on Brazilian federal highways by specifically isolating human-factor variables. Leveraging an extensive dataset from the Brazilian Federal Highway Police (2013–2025), we conducted a comparative analysis of three ensemble-based algorithms: Random Forest, XGBoost, and LightGBM. To ensure model stability and generalization, hyperparameter optimization was executed using RandomizedSearchCV with five-fold cross-validation. The experimental results demonstrate that while Random Forest achieved high training accuracy, LightGBM emerged as the superior model for safety-critical deployment, achieving a balanced ROC-AUC of 0.843 and a superior recall, effectively minimizing life-threatening false negatives. Furthermore, this research integrates SHAP (SHapley Additive exPlanations) to provide a knowledge-based interpretation of the model's decisions. The XAI analysis reveals that “Accident Type” and specific “Human Factor Categories” are the most significant predictors of fatality. The findings provide a transparent, data-driven decision support tool for transportation authorities to implement targeted interventions, bridging the gap between complex black-box models and actionable road safety knowledge.
Power transmission network plays a crucial role in maintaining the stability and reliability of electrical systems. Fault detection and predictive maintenance are essential to ensure continuous operation and minimize downtimes, but traditional fault detection methods face challenges, particularly in remote areas where manual inspections are impractical. This paper presents a framework to enhance the steadiness and efficacy of power transmission networks through advanced fault detection and predictive maintenance. The proposed framework begins with data collection from power transmission sensors, including voltage, current, and temperature readings, along with historical fault records. Next, data pre-processing is performed using median imputation to handle missing values and categorical encoding to transform non-numeric data into numerical form. Feature extraction follows, where time-domain features like Peak-to-Peak Value, RMS, and Zero-Crossing Rate are computed to detect potential faults. The CatBoost model is then trained on the extracted features, and hyperparameter optimization is conducted using the Coati Optimization Algorithm. Once trained, the model performs fault detection and prediction, identifying faults such as Transformer Failures, Overheating, and Line Breakages. The model is assessed using metrics like accuracy of 99.42%, precision of 99.37%, recall of 99.40%, and F1-score of 99.38%. The framework achieves high performance in detecting faults and can be deployed in power transmission systems for proactive maintenance, reducing reliance on manual inspections, improving system reliability, and addressing challenges in remote locations.
Text summarization systems often struggle with selecting salient content, avoiding repetition, and handling out-of-vocabulary entities. We address these issues with a two-stage approach: a supervised sentence-ranking head (SRM-head) first selects the top-N sentences, and a Transformer generator then produces the summary. The generator is augmented with a time penalty in encoder-decoder attention to discourage reattending to recently focused source positions, and with a pointer mechanism that copies salient spans, thereby improving entity and number fidelity. Experiments on CNN/DailyMail and WikiHow, plus an additional evaluation on XSum, show that our model attains competitive ROUGE scores against recent pretrained systems while using lightweight, modular components.
With the rapid development of maritime commerce, the complexity of ship operation environments continues to grow, so that it is crutial for ship designers to make accurate forecasting of ship performance under various sea conditions. However, there exist the following challenges in ship performance evaluation and optimization. The first challenge is the computational overhead and slow response of numerical simulations based on Computational Fluid Dynamics (CFD), and the second is the inability of traditional machine learning to achieve the computational accuracy of numerical simulations. To address these issues, this paper proposes an online surrogate model, called Increformer, for ship performance prediction. The Increformer leverages continuous self-attention mechanisms to explore the temporal dependencies between feature variables and employs continuous normalization mechanisms to handle non-stationary data issues. In addition, in order to improve prediction accuracy by the model, we employ an incremental training strategy based on elastic weight consolidation to acquire new knowledge from data streams. Experiments are conducted by using historical performance data from various types of vessels including KCS, Wigley-III, and C60. The results demonstrate that the Increformer model effectively captures the temporal information and inter-dimensional correlations in the data, with the accuracy of ship performance prediction enhanced significantly. Furthermore, ablation experiments are also carried out to assess the effectiveness and necessity of the continuous normalization mechanism, continuous attention mechanism, and incremental training strategy for the Increformer model. The findings validate the accuracy and universality of the proposed model. It is also shown that the Increformer model adeptly captures trends and fluctuations in ship sequence data and thus provides a reliable solution for ship performance prediction.
Aspect-based dialogue sentiment quadruple analysis (DiaASQ) is a critical task in sentiment analysis, aiming to extract sentiment quadruples (target, aspect, opinion, sentiment polarity) from dialogues. Existing methods primarily focus on single-sentence sentiment analysis, often neglecting the rich contextual information and long-range dependencies in multi-turn dialogues. To address this limitation, we propose a novel memory framework, Memory, which incorporates adaptive contextual memory mechanisms to simulate human-like emotional refinement during conversations. Our framework consists of three key components: a Contextual Knowledge Memorizer to capture token-level syntactic-semantic dependencies, an Utterance-level Sentiment Interactor to model speaker-respondent dynamics, and a Multi-granularity Memory Integrator to fuse token-level and utterance-level information for precise sentiment relationship extraction. Extensive experiments on two benchmark datasets demonstrate the framework’s superiority, achieving 10.14% and 6.03% improvements in Micro-F1, and 13.07% and 5.60% improvements in Iden-F1 on Chinese and English datasets, respectively.
The core challenge of CTR prediction lies in how to effectively model feature semantics and their complex nonlinear interactions from high-dimensional sparse data. Existing methods mostly adopt static embedding in feature representation, ignoring the contextual dependency of semantics. In interaction modeling, traditional MLP suffers from gradient forgetting and low parameter efficiency, and are difficult to efficiently fit complex patterns such as non monotonicity. In response to the aforementioned limitations, this study proposes EFMNet, a hybrid architecture that integrates dynamic representation learning and frequency domain function approximation. Firstly, a feature dynamic enhancement module is designed, which generates context-aware complementary features through FRNet to reconstruct semantics, and uses SENet to evaluate the contribution at the channel level. Through affine transformation, it generates sample adaptive dynamic embedding. On this basis, an interaction paradigm of implicit and explicit collaboration is constructed: Fourier KAN is introduced as the implicit backbone, and its learnable edge functions based on Fourier basis are utilized to fit the nonlinear dependencies among features with high parameter efficiency; at the same time, CrossNet is integrated to explicitly model higher-order cross correlations, ensuring the effective transmission of key combined signals. Experiments on the Criteo and Avazu benchmarks show that EFMNet performs obviously better than strong baselines such as EulerNet and FinalMLP (e.g Criteo_AUC: 0.8164, Criteo_Logloss: 0.4427, Avazu_AUC: 0.7881, Avazu_Logloss: 0.3684), and its efficiency is comparable to that of the widely deployed DeepFM in the industry. This work validates the effectiveness of dynamic feature modeling and frequency domain approximation in CTR prediction, providing new ideas for recommendation system design.
Algerian Arabic (Darija) dominates digital communication in North Africa yet remains severely under-resourced in Natural Language Processing (NLP), hindering the development of robust applications for social media analysis and e-commerce. This paper addresses this scarcity by presenting a systematic framework for constructing and benchmarking Named Entity Recognition (NER) and Entity Linking (EL) resources tailored to the dialect’s linguistic complexity. We introduce a large-scale, multi-script dataset constructed through a novel hybrid methodology that integrates manual annotation of authentic texts, automated knowledge graph extraction from Wikidata, and rule-based synthetic generation. This approach ensures diverse coverage across ten semantic categories while explicitly addressing the challenges of code-switching and orthographic variation (Arabizi and Arabic script). A transformer-based model (XLM-RoBERTa) fine-tuned on this resource achieves state-of-the-art performance, demonstrating significant robustness compared to existing baselines. Beyond the dataset, we provide a practical deployment interface and comprehensive evaluation metrics, establishing a crucial foundation for advancing NLP capabilities in North African dialects and facilitating downstream tasks such as content moderation and cultural heritage preservation.
