Search Results (216)

Accurate recognition of estrus behavior in sows is of great importance for achieving scientific breeding management, improving reproductive efficiency, and reducing labor costs in modern pig farms. However, due to the evident spatiotemporal continuity, stage-specific changes, and ambiguous category boundaries of estrus behaviors, traditional methods based on static images or manual observation suffer from low efficiency and high misjudgment rates in practical applications. To address these issues, this study follows a video-based behavior recognition approach and designs three deep learning model structures: (Convolutional Neural Network combined with Long Short-Term Memory) CNN + LSTM, (Three-Dimensional Convolutional Neural Network) 3D-CNN, and (Convolutional Neural Network combined with Temporal Convolutional Network) CNN + TCN, aiming to achieve high-precision recognition and classification of four key behaviors (SOB, SOC, SOS, SOW) during the estrus process in sows. In terms of data processing, a sliding window strategy was adopted to slice the annotated video sequences, constructing image sequence samples with uniform length. The training, validation, and test sets were divided in a 6:2:2 ratio, ensuring balanced distribution of behavior categories. During model training and evaluation, a systematic comparative analysis was conducted from multiple aspects, including loss function variation (Loss), accuracy, precision, recall, F1-score, confusion matrix, and ROC-AUC curves. Experimental results show that the CNN + TCN model performed best overall, with validation accuracy exceeding 0.98, F1-score approaching 1.0, and an average AUC value of 0.9988, demonstrating excellent recognition accuracy and generalization ability. The 3D-CNN model performed well in recognizing short-term dynamic behaviors (such as SOC), achieving a validation F1-score of 0.91 and an AUC of 0.770, making it suitable for high-frequency, short-duration behavior recognition. The CNN + LSTM model exhibited good robustness in handling long-duration static behaviors (such as SOB and SOS), with a validation accuracy of 0.99 and an AUC of 0.9965. In addition, this study further developed an intelligent recognition system with front-end visualization, result feedback, and user interaction functions, enabling local deployment and real-time application of the model in farming environments, thus providing practical technical support for the digitalization and intelligentization of reproductive management in large-scale pig farms. Full article

Online advertising is vital for reaching target audiences and promoting products. In 2020, US online advertising revenue increased by 12.2% to $139.8 billion. The industry is projected to reach $487.32 billion by 2030. Artificial intelligence has improved click-through rates (CTR), enabling personalized advertising content by analyzing user behavior and providing real-time predictions. This review examines the latest CTR prediction solutions, particularly those based on deep learning, over the past three years. This timeframe was chosen because CTR prediction has rapidly advanced in recent years, particularly with transformer architectures, multimodal fusion techniques, and industrial applications. By focusing on the last three years, the review highlights the most relevant developments not covered in earlier surveys. This review classifies CTR prediction methods into two main categories: CTR prediction techniques employing text and CTR prediction approaches utilizing multivariate data. The methods that use multivariate data to predict CTR are further categorized into four classes: graph-based methods, feature-interaction-based techniques, customer-behavior approaches, and cross-domain methods. The review also outlines current challenges and future research opportunities. The review highlights that graph-based and multimodal methods currently dominate state-of-the-art CTR prediction, while feature-interaction and cross-domain approaches provide complementary strengths. These key takeaways frame open challenges and emerging research directions. Full article

Electricity theft in power grids results in significant economic losses for utility companies. While machine learning (ML) methods have shown promising results in detecting such frauds, they often suffer from low detection rates, leading to excessive physical inspections. In this study, we attempted to solve the above-mentioned problem using a novel approach. The proposed framework utilizes the intelligence of Siamese network architecture with the Triplet Loss function to detect electricity theft using a labeled dataset obtained from Multan Electric Power Company (MEPCO), Pakistan. The proposed method involves analyzing and comparing the consumption patterns of honest and fraudulent consumers, enabling the model to distinguish between the two categories with enhanced accuracy and detection rates. We incorporate advanced feature extraction techniques and data mining methods to transform raw consumption data into informative features, such as time-based consumption profiles and anomalous load behaviors, which are crucial for detecting abnormal patterns in electricity consumption. The refined dataset is then used to train the Siamese network, where the Triplet Loss function optimizes the model by maximizing the distance between dissimilar (fraudulent and honest) consumption patterns while minimizing the distance among similar ones. The results demonstrate that our proposed solution outperforms traditional methods by significantly improving accuracy (95.4%) and precision (92%). Eventually, the integration of feature extraction with Siamese networks and Triplet Loss offers a scalable and robust framework for enhancing the security and operational efficiency of power grids. Full article

(1) Background: Soccer action recognition (SAR) is essential in modern sports analytics, supporting automated performance evaluation, tactical strategy analysis, and detailed player behavior modeling. Although recent advances in deep learning and computer vision have enhanced SAR capabilities, many existing methods remain limited to coarse-grained classifications, grouping actions into broad categories such as attacking, defending, or goalkeeping. These models often fall short in capturing fine-grained distinctions, contextual nuances, and long-range temporal dependencies. Transformer-based approaches offer potential improvements but are typically constrained by the need for large-scale datasets and high computational demands, limiting their practical applicability. Moreover, current SAR systems frequently encounter difficulties in handling occlusions, background clutter, and variable camera angles, which contribute to misclassifications and reduced accuracy. (2) Methods: To overcome these challenges, we propose TAT-SARNet, a structured framework designed for accurate and fine-grained SAR. The model begins by applying Sparse Dilated Attention (SDA) to emphasize relevant spatial dependencies while mitigating background noise. Refined spatial features are then processed through the Split-Stream Feature Processing Module (SSFPM), which separately extracts appearance-based (RGB) and motion-based (optical flow) features using ResNet and 3D CNNs. These features are temporally refined by the Multi-Granular Temporal Processing (MGTP) module, which integrates ResIncept Patch Consolidation (RIPC) and Progressive Scale Construction Module (PSCM) to capture both short- and long-range temporal patterns. The output is then fused via the Context-Guided Dual Transformer (CGDT), which models spatiotemporal interactions through a Bi-Transformer Connector (BTC) and Channel–Spatial Attention Block (CSAB); (3) Results: Finally, the Cascaded Temporal Classification (CTC) module maps these features to fine-grained action categories, enabling robust recognition even under challenging conditions such as occlusions and rapid movements. (4) Conclusions: This end-to-end architecture ensures high precision in complex real-world soccer scenarios. Full article

Next-point-of-interest (POI) recommendation aims to model user preferences based on historical information to predict future mobility behavior, which has significant application value in fields such as urban planning, traffic management, and optimizing business decisions. However, existing methods often overlook the differences in location, time, and category information features, fail to fully utilize information from various modalities, and lack effective solutions for addressing users’ incidental behavior. Additionally, existing methods are somewhat lacking in capturing users’ personalized preferences. To address these issues, we propose a new method called Multi-Branch Spectral Network with Contrastive Learning (MBSCL-Net) for next-POI recommendation. We use a multihead attention mechanism to separately capture the distinct features of location, time, and category information, and then fuse the captured features to effectively integrate cross-modal features, avoid feature confusion, and achieve effective modeling of multi-modal information. We propose converting the time-domain information of user check-ins into frequency-domain information through Fourier transformation, directly enhancing the low-frequency signals of users’ periodic behavior and suppressing occasional high-frequency noise, thereby greatly alleviating noise interference caused by the introduction of too much information. Additionally, we introduced contrastive learning loss to distinguish user behavior patterns and better model personalized preferences. Extensive experiments on two real-world datasets demonstrate that MBSCL-Net outperforms state-of-the-art (SOTA) methods. Full article

The rise of generative AI in higher education has disrupted our traditional understandings of academic integrity, moving our focus from clear-cut infractions to evolving ethical judgment. In this study, a survey of 401 students from major U.S. universities provides insight into how beliefs, behaviors, and policy awareness intersect in shaping how students interact with AI-assisted writing. The findings indicate that students’ ethical beliefs—not institutional policies—are the strongest predictors of perceived misconduct and actual AI use in writing. Policy awareness was found to have no significant effect on ethical judgments or behavior. Instead, students who believe AI writing is cheating were found to be substantially less likely to view it as ethical or engage with it. These findings suggest that many students do not treat AI use in learning activities as an extension of conventional cheating (e.g., plagiarism), but rather as a distinct category of academic conduct/misconduct. Rather than using punitive models to attempt to punish students for using AI, this study suggests that education about AI ethics and the risk of AI overreliance may prove more successful for curbing unethical AI use in higher education. Full article

Conglomerate reservoirs present significant technical challenges during drilling operations due to their complex mineral composition and heterogeneous characteristics, yet the quantitative relationships between mineral composition and microscopic mechanical behavior remain poorly understood. To elucidate the variation patterns of conglomerate micromechanical properties and their mineralogical control mechanisms, this study develops a novel multi-scale characterization methodology. This approach uniquely couples nanoindentation technology, micro-zone X-ray diffraction analysis, and machine learning algorithms to systematically investigate micromechanical properties of conglomerate samples from different regions. Hierarchical clustering algorithms successfully classified conglomerate micro-regions into three lithofacies categories with distinct mechanical differences: hard (elastic modulus: 81.90 GPa, hardness: 7.83 GPa), medium-hard (elastic modulus: 54.97 GPa, hardness: 3.87 GPa), and soft lithofacies (elastic modulus: 25.21 GPa, hardness: 1.15 GPa). Correlation analysis reveals that quartz (SiO₂) content shows significant positive correlation with elastic modulus (r = 0.52) and hardness (r = 0.51), while clay minerals (r = −0.37) and plagioclase content (r = −0.48) exhibit negative correlations with elastic modulus. Mineral phase spatial distribution patterns control the heterogeneous characteristics of conglomerate micromechanical properties. Additionally, a random forest regression model successfully predicts mineral content based on hardness and elastic modulus measurements with high accuracy. These findings bridge the gap between microscopic mineral properties and macroscopic drilling performance, enabling real-time formation strength assessment and providing scientific foundation for optimizing drilling strategies in heterogeneous conglomerate formations. Full article

Background: The human visual system is capable of processing large quantities of visual objects with varying levels of abstraction. The brain also exhibits hierarchical integration and learning capabilities that combine various attributes of visual objects (e.g., color, shape, local features, and categories) into coherent representations. However, prevailing theories in visual neuroscience employ simple stimuli or natural images with uncontrolled feature correlations, which constrains the systematic investigation of multidimensional representation dynamics. Methods: In this study, we aimed to bridge this methodological gap by developing a novel large tangram paradigm in visual cognition research and proposing cognitive-associative encoding as a mathematical basis. Critical representation dimensions—including animacy, abstraction level, and local feature density—were computed across a public dataset of over 900 tangrams, enabling the construction of a hierarchical model of visual representation. Results: Neural responses to 85 representative images were recorded using Electroencephalography (n = 24), and subsequent behavioral analyses and neural decoding revealed that distinct representational dimensions are independently encoded and dynamically expressed at different stages of cognitive processing. Furthermore, representational similarity analysis and temporal generalization analysis indicated that higher-order cognitive processes, such as “change of mind,” reflect the selective activation or suppression of local feature processing. Conclusions: These findings demonstrate that tangram stimuli, structured through cognitive-associative encoding, provide a generalizable computational framework for investigating the dynamic stages of human visual object cognition. Full article

Background: Studies have shown that humans can rapidly learn the shape of new objects or adjust their behavior when encountering novel situations. Research on visual cognition in the brain further indicates that the ventral visual pathway plays a critical role in core object recognition. While existing studies often focus on microscopic simulations of individual neural structures, few adopt a holistic, system-level perspective, making it difficult to achieve robust few-shot learning capabilities. Method: Inspired by the mechanisms and processes of the ventral visual stream, this paper proposes a computational model with a macroscopic neural architecture for few-shot learning. We reproduce the feature extraction functions of V1 and V2 using a well-trained Vision Transformer (ViT) and model the neuronal activity in V4 and IT using two neural fields. By connecting these neurons based on Hebbian learning rules, the proposed model stores the feature and category information of the input samples during support training. Results: By employing a scale adaptation strategy, the proposed model emulates visual neural mechanisms, enables efficient learning, and outperforms state-of-the-art few-shot learning algorithms in comparative experiments on real-world image datasets, demonstrating human-like learning capabilities. Conclusion: Experimental results demonstrate that our ventral-stream-inspired machine-learning model achieves effective few-shot learning on real-world datasets. Full article

Depression is a serious mental health disorder affecting cognition, emotions, and behavior. It impacts over 300 million people globally, with mental health care costs exceeding $1 trillion annually. Traditional diagnostic methods are often expensive, time-consuming, stigmatizing, and difficult to access. This study leverages NLP techniques to identify depressive cues in social media posts, focusing on both standard Urdu and code-mixed Roman Urdu, which are often overlooked in existing research. To the best of our knowledge, a script-conversion and combination-based approach for Roman Urdu and Nastaliq Urdu has not been explored earlier. To address this gap, our study makes four key contributions. First, we created a manually annotated dataset named Ruda-2025, containing posts in code-mixed Roman Urdu and Nastaliq Urdu for both binary and multiclass classification. The binary classes are depression” and not depression, with the depression class further divided into fine-grained categories: Mild, Moderate, and Severe depression alongside not depression. Second, we applied first-time two novel techniques to the RUDA-2025 dataset: (1) script-conversion approach that translates between code-mixed Roman Urdu and Standard Urdu and (2) combination-based approach that merges both scripts to make a single dataset to address linguistic challenges in depression assessment. Finally, we employed 60 different experiments using a combination of traditional machine learning and deep learning techniques to find the best-fit model for the detection of mental disorder. Based on our analysis, our proposed model (mBERT) using custom attention mechanism outperformed baseline (XGB) in combination-based, code-mixed Roman and Nastaliq Urdu script conversions. Full article

Interactions between human behavior, legal regulations, and monitoring technology in road traffic systems provide an everyday example of complex biosocial–technical systems. In this paper, a study is reported that investigated the potential for a thrifty world model to predict consequences from choices about road traffic system design. Colloquially, the term thrifty means economical. In physics, the term thrifty is related to the principle of least action. Predictions were made with algebraic machine learning, which combines predefined embeddings with ongoing learning from data. The thrifty world model comprises three categories that encompass a total of only eight system design choice options. Results indicate that the thrifty world model is sufficient to encompass biosocial–technical complexity in predictions of where and when it is most likely that accidents will occur. Overall, it is argued that thrifty world models can provide a practical alternative to large photo-realistic world models, which can contribute to explainable artificial intelligence (AI) and to frugal AI. Full article

Aim: To explore the experiences, beliefs, and values of patients who participated in a two-arm randomized clinical trial assessing a nurse-led intervention program for chronic pain self-management, which demonstrated positive effects on pain reduction, depression, and anxiety, and on health-related quality of life 24 months after completion of the program. Design: Descriptive phenomenological qualitative study. Methods: Patients were recruited via telephone, informed about the study, and invited to participate in an individual interview at a place of their choice (hospital or home). All interviews were audiotaped, and an inductive thematic analysis was performed. Results: Seven interviews were carried out between both groups. Six emerging categories were found: effective relationship with the healthcare system, learning to live with pain, family and social support, behaviors regarding pain, resources for self-management, and concomitant determinants. Conclusions: Patients report key aspects that help us to understand the impact of this type of nurse-led group intervention: the intrinsic therapeutic effect of participating in the program itself, the ability to learn to live with pain, the importance of family and social support, the modification of pain-related behaviors, and the identification of resources for self-care. The findings highlight the need for gender-sensitive, individualized care approaches to chronic pain, addressing stigma and social context. Expanding community-based programs and supporting caregivers is essential, as is further research into gender roles, family dynamics, and work-related factors. Full article

In recent decades, university dormitories have gradually evolved from traditional residential spaces into educationally meaningful venues that support informal learning. However, limited research has explored how supportive environmental factors within dormitories influence students’ informal learning experiences. This study aims to evaluate key environmental factors that affect students’ satisfaction with informal learning in dormitory settings. Based on a comprehensive literature review, two types of informal learning behaviors—individual and collaborative—were defined, and a multi-dimensional evaluation framework comprising five categories and 26 environmental indicators was established. Field observations and structured questionnaires were employed to assess students’ satisfaction with each environmental factor and their overall informal learning experiences. Quantitative analyses were conducted to examine the relationships between environmental conditions and learning satisfaction. Results show that all five-factor categories—spatial designs, natural environments, physical settings, social aspects, and resources—positively influence informal learning, with resources being the most impactful. While environmental influences on individual and collaborative learning exhibit minor differences, the overall patterns are consistent. Compared to other informal learning spaces on campus, dormitory users place greater emphasis on spatial controllability. This study further demonstrates the cognitive and emotional value of dormitory environments and proposes targeted directions for optimizing them as informal learning spaces. Full article

Employees’ green voice behavior (GVB), as a specific category of extra-role green behavior, plays a vital role in promoting a firm’s sustainable development. However, its underlying mechanism has not been sufficiently explored. Drawing on social learning theory (SLT), this study proposes a research model that examines the indirect influence of environmentally-specific transformational leadership (ESTFL) on GVB via psychological empowerment (PE) and ecological reflexivity (ER) as well as the moderating role of person-supervisor value congruence (PSVC). To achieve the research goals, we conducted a two-wave online survey via the convenience sampling method to collect data from 530 employees and 106 direct supervisors working in the manufacturing, hospitality and service, energy production, construction, transportation, information and communication, and finance industries in China. Regression analyses and CFA based on SPSS and Mplus were employed to test and validate the research model. Our findings show that PE and ER both partially mediated the positive association between ESTFL and GVB. Moreover, PSVC moderated the mediating effects of ESTFL on GVB via PE and ER. This study advances empirical research regarding how leadership impacts GVB by revealing dual cognitive mechanisms and identifying its boundary condition. It also offers managerial implications for leaders and enterprises in China to promote employees’ GVB and improve sustainable management. Full article

Software requirements are primarily classified into functional and non-functional requirements. While research has explored automated multiclass classification of non-functional requirements, functional requirements remain largely unexplored. This study addressed that gap by introducing a comprehensive dataset comprising 9529 functional requirements from 315 diverse projects. The requirements are classified into five categories: ubiquitous, event-driven, state-driven, unwanted behavior, and optional capabilities. Natural Language Processing (NLP), machine learning (ML), and deep learning (DL) techniques are employed to enable automated classification. All software requirements underwent several procedures, including normalization and feature extraction techniques such as TF-IDF. A series of Machine learning (ML) and deep learning (DL) experiments were conducted to classify subcategories of functional requirements. Among the trained models, the convolutional neural network achieved the highest performance, with an accuracy of 93, followed by the long short-term memory network with an accuracy of 92, outperforming traditional decision-tree-based methods. This work offers a foundation for precise requirement classification tools by providing both the dataset and an automated classification approach. Full article