Recently, online education and video-conferencing platforms have experienced substantial growth. Owing to the limited scope of virtual classrooms, it is difficult for instructors to assess learners’ attention and comprehension in real time during instruction. In digital learning environments, an automated feedback mechanism that informs instructors about learners’ attentiveness at any given time would be beneficial. To this end, we present a novel computer vision-based approach for analyzing learners’ affective states and quantifying their attentiveness. The foundation of our approach is a multioutput classification model trained on the publicly available DAiSEE dataset, designed to accurately identify learning-centered affective states such as boredom, engagement, confusion, and frustration. A key contribution of this work is the novel attentiveness index, which is a composite metric derived from the four affective states to quantify perceived attentiveness. Furthermore, we present a complete end-to-end pipeline that integrates this index into a practical, real-time system, processing learners’ live video feeds to provide instructors with detailed, actionable analytics. The robustness of this approach is validated by experimental outcomes in which our underlying models demonstrate state-of-the-art performance in affective state classification, thereby supporting the reliability of the entire system. The result is a comprehensive and practical real-time solution that is both deployable and easy to use, offering instructors an effective tool for gauging learner attentiveness in virtual classrooms.

The rise of e-learning has transformed instructional delivery, improving accessibility and flexibility. However, it presents challenges in maintaining real-time engagement and providing responsive support, as instructors lack access to students’ subtle behavioral cues, like facial expressions or affective shifts, that are readily observed in physical classrooms. This further limits timely and dynamic instructional adaptation. In this work, we present an end-to-end system that addresses these challenges by integrating real-time attentiveness detection with adaptive feedback generation. Our approach uses a Convolutional Neural Network (CNN) to classify key affective states (boredom, engagement, confusion, and frustration) based on students’ facial expressions. These classifications are modeled into an attentiveness index, which informs a heuristic-based prompting framework that dynamically invokes a Large Language Model (LLM) to generate context-aware feedback. The LLM produces adaptive suggestions for instructors to adjust delivery style, pacing, or content, while simultaneously providing students with personalized explanations and curated learning resources aligned with the inattentive lecture segments. Our experimental results demonstrate strong classification performance for affective states compared to existing deep learning approaches. By combining real-time affect recognition with adaptive feedback, our system bridges detection and intervention, offering a scalable solution that supports both instructors and learners in enhancing engagement and comprehension.

Deep learning has revolutionized medical image processing, offering powerful tools for diagnosing critical conditions like brain tumors, skin cancer, and pneumonia. However, the opaque nature of these models necessitates interpretability, especially in healthcare applications where transparency and trust are vital. This study addresses this challenge by integrating explainable artificial intelligence (XAI) techniques, specifically local interpretable model-agnostic explanations (LIME) and gradient-weighted class activation mapping (Grad-CAM), to enhance model transparency. A comparative analysis of five deep learning models-CNN, XceptionNet, EfficientNet, VGG, and ResNet-is conducted across three distinct medical image classification tasks. Performance is evaluated using accuracy, precision, recall, F1 score, specificity, and AUC score, while LIME and Grad-CAM provide visual insights into the regions influencing model predictions, supporting clinical validation. The findings underscore the importance of XAI in improving model transparency, offering valuable insights into the reliability and applicability of deep learning models in medical image analysis, ultimately aiding in the development of more explainable and trustworthy AI systems for healthcare.

Question answering (QA) portals like Quora, Stack Overflow, AskUbuntu, Yahoo! Answers, Reddit, and Wiki Answers have emerged as hubs of curiosity, highlighting the rising demands for easily accessible information and are drawing focus to hundreds of millions of questions. The efficient utilization of these questions and associated answers has become significantly vital for these QA websites. The similarity-based information retrieval methods provide a ranked list of potentially relevant questions, and the users have to spend significant time sifting through the results to discover the best answer. This paper aims to provide a precise, comprehensive, summarized answer to the user asked query using extracted keywords that offer valuable insights into relevant content. The research work presents a Query focused Answer Summarization framework using Keyword Extraction (QFAS-KE). It is a four-stage framework, including query question pre-processing, semantic question search (utilizing SBERT and FAISS vector database), answer retrieval and re-ranking (utilizing BERT-based bi-encoder and cross-encoder), and answer summary generation (using fine-tuned transformers such as BART, PEGASUS, T5) with keyword guidance (using a keyword extractor such as KeyBERT). The results conceptualize the efficacy of the proposed framework on task-specific datasets (CNN/DailyMail and MS-MARCO) over the ROUGE metric. The model outperformed existing baseline models on CNN/DailyMail dataset with a value of 47.5 (PEGASUS), 46.2 (BART), and 45.1 (T5) in terms of ROUGE-1 and on MS-MARCO dataset with a value of 75.18 (PEGASUS), 79.02 (BART), and 74.69 (T5) in terms of ROUGE-L.

This study proposes an ensemble approach to develop a tomato ripeness and shelf life prediction system based on defects and color intensity. The dataset has been created by designing an image acquisition system to capture 3450 images. The image outliers are overcome by various image processing steps. In the context of our proposed ensemble technique, we employ a variety of expert-based regressors, which encompass SVM, DT, RF, and GBM, in order to ensure distinctiveness. When assessing tomato characteristics for the purpose of quality determination, it becomes crucial to take into account attributes such as size, color, shape, texture, taste, nutritional content, defects, and ripeness, among others. A total of 13 features are manually extracted, whereas the Inception V3 model autonomously generates 2048 automated features. These 2048 features are then streamlined through the application of PCA dimensionality reduction, resulting in a final selection of 50 automated features. Altogether, the proposed work leverages a total of 63 features. The output of the ripeness regression models is divided into three classes based on their ripeness index and color magnitude to get the shelf-life of the tomatoes as Store, Sell, and Discount. The stacking technique is employed to achieve a final prediction of tomato shelf life with an impressive accuracy rate of 90.35%. These findings highlight that incorporating a variety of features, diverse pre-processing techniques, and proficient machine learning regressors can introduce substantial diversity in the ensemble approach, leading to enhanced accuracy when compared to conventional machine learning models. The proposed model has been rigorously compared with numerous state-of-the-art detection methods, yielding highly promising results.

In recent times, online education and the usage of video-conferencing platforms have experienced massive growth. Due to the limited scope of a virtual classroom, it may become difficult for instructors to analyze learners' attention and comprehension in real time while teaching. In the digital mode of education, it would be beneficial for instructors to have an automated feedback mechanism to be informed regarding learners' attentiveness at any given time. This research presents a novel computer vision-based approach to analyze and quantify learners' attentiveness, engagement, and other affective states within online learning scenarios. This work presents the development of a multiclass multioutput classification method using convolutional neural networks on a publicly available dataset - DAiSEE. A machine learning-based algorithm is developed on top of the classification model that outputs a comprehensive attentiveness index of the learners. Furthermore, an end-to-end pipeline is proposed through which learners' live video feed is processed, providing detailed attentiveness analytics of the learners to the instructors. By comparing the experimental outcomes of the proposed method against those of previous methods, it is demonstrated that the proposed method exhibits better attentiveness detection than state-of-the-art methods. The proposed system is a comprehensive, practical, and real-time solution that is deployable and easy to use. The experimental results also demonstrate the system's efficiency in gauging learners' attentiveness.

Identifying offensive and discriminatory content, commonly referred to as hate speech, within textual data is a critical task. This study addresses the task of identifying hate speech in textual data, focusing on the challenge of selecting optimal word embedding methods and classifiers. Leveraging the Google Jigsaw dataset, the research employs explainable artificial intelligence (XAI) for hate speech detection. Following preprocessing, which includes converting text to lowercase, removing punctuation, extra whitespace, numbers, and non-ASCII characters, a thorough analysis reveals high-frequency words. The research extensively compares three-word embedding techniques—CountVectorizer, GloVe, and bidirectional encoder representations from transformers (BERT)—in combination with two machine learning models (support vector classifier and logistic regression) and four deep learning models [artificial neural network (ANN), recurrent neural network (RNN), bidirectional gated recurrent unit (Bi-GRU), bidirectional long-short term memory (Bi-LSTM)] for hate speech detection. The fusion of BERT with a bidirectional gated recurrent unit (Bi-GRU) achieved an impressive accuracy of 92

In the era of Online Social Networks (OSNs), efficient communication is crucial for disseminating information to audiences swiftly. However, this convenience comes with challenges as Clickbait and malicious activities pose significant risks to users and such cybercrimes result in substantial financial losses annually. Addressing this issue is paramount to safeguard users and maintain the integrity of OSNs. This study proposes a two-level Uniform Resource Locator(URL) detection technique to combat these threats. Firstly, a human-annotated database is created from various phishing websites, facilitating the differentiation between authentic and fraudulent URLs. Then, a deep Long-Short Term Memory (LSTM) neural network, leveraging a 100-dimensional Glove word embedding at the word level, extracts feature vectors for URLs which enables accurate predictions regarding URL authenticity. Additionally, a ground dataset is established, categorizing URLs as genuine or fake based on trustworthy websites.

In response to the critical issue of milk adulteration jeopardizing both the nutritional integrity of milk and the health of consumers, this paper presents an innovative Artificial Intelligence (AI) enabled Internet of Things (IoT) based multi-sensor system. The escalating consumption of milk as a pivotal nutritional source necessitates robust measures to ensure its safety and quality. Traditional methods of detecting adulteration have shown limitations, prompting the development of an automated and advanced approach. The proposed system integrates various sensors capable of real-time measurement, including pH, electrical conductivity (EC), temperature, gas parameters, and Volatile Organic Compounds (VOC) parameters. This comprehensive approach extends to measuring key constituents of milk samples like Fat, Protein, Solids Not Fat (SNF), Lactose, and Gravity values. To address specific adulterants—Urea, Starch, Sodium Bicarbonate, Maltodextrin, and Formaldehyde—a machine learning-based ensemble technique is employed for classification. This ensemble method outperforms conventional algorithms like RF, Light GBM, and Extra Trees Classifiers, achieving an impressive 96% accuracy rate in detecting adulterants within the milk dataset. The pivotal contribution of this study lies in the development of an IoT-based data acquisition device that seamlessly integrates with the sensor system, enabling efficient and precise measurements. Additionally, XAI is used to analyse the results obtained by the proposed model. For this, a framework called SHAP (SHapley Additive exPlanations) analysis is employed to elucidate the decision-making process of the ensemble model, enhancing the interpretability of results. By virtue of its real-time monitoring capabilities and accurate classification, the AI-enabled IoT-based multi-sensor system emerges as a promising solution for addressing milk adulteration. This innovation holds the potential to bolster milk quality control measures in the dairy industry. The system's ability to swiftly detect and categorize adulterants underscores its significance in combating the pervasive issue of compromised milk quality, thereby ensuring consumer safety and fostering industry integrity.

The COVID-19 pandemic has significantly shifted how people learn, with virtual learning becoming the norm. Students no longer need to be physically present in a specific location to gain knowledge. This paradigm shift has led to new challenges, particularly regarding instructor feedback during e-learning sessions. Unlike traditional learning, where face-to-face interaction allows instructors to give feedback based on student’s learning behaviour, the lack of such interaction in e-learning poses a challenge. To address this, two methods: a computer vision-based and EEG-based, are used in the present work to monitor a student’s cognitive behaviour during e-learning sessions. The former involves using a webcam to capture facial cues, such as facial expressions, eye blinks, and head movements, while the latter requires using an EEG device, such as the Neurosky wireless device, to record EEG waves. By collecting and analysing this data, teachers can identify students struggling in the e-learning environment. This paper explores using facial cues and EEG signals to predict a student’s cognitive behaviour in the form of attentiveness or inattentiveness throughout the e-learning sessions. The experimental results achieved an accuracy of 92.33% using facial cues, while the accuracy achieved using EEG was 91.68%. Overall, efficient detection of students’ learning situations can help teachers to create a more responsive and effective e-learning environment that meets the needs of all students.

Automatic question generation and automatic question answering from text is a fundamental academic tool that serves a wide range of purposes, including self-study, coursework, educational assessment, and many more. Manual construction of questions is a time-consuming and complicated process that requires experience, whereas automating the process diminishes the costs of manual question creation and fulfills the need for a persistent supply of questions for the tutors and self-evaluators. This paper uses an encoder–decoder architecture-based text-to-text transfer transformer (T5) intending to generate several types of question–answer pairs over a given context, including subjective question–answers having short and long answers, fill-in-the-blanks-type question–answers, Boolean answer (yes-or-no)-type questions, and multiple-choice question-answers. The model has been evaluated on benchmark datasets—SQuAD, QuAC, and BoolQ; over automated metrics—BLEU, ROUGE, METEOR, F1, and accuracy. The model outperformed previous baseline models, with 18.87 and 25.24 scores over BLEU-4 and METEOR metrics, respectively. The paper also demonstrates that the proposed system efficiently generates question–answer pairs where the baseline approaches struggled. The evaluation analysis also shows that the generated question–answer pairs are comparable with existing systems and even better in terms of diversity. Also, the generated questions are grammatically and contextually correct, and the answer generated matches the question in the textual context.

This paper presents a method for unsupervised classification of entities by a group of agents with unknown domains and levels of expertise. In contrast to the existing methods based on majority voting (“wisdom of the crowd”) and their extensions by expectation-maximization procedures, the suggested method first determines the levels of the agents’ expertise and then weights their opinions by their expertise level. In particular, we assume that agents will have relatively closer classifications in their field of expertise. Therefore, the expert agents are recognized by using a weighted Hamming distance between their classifications, and then the final classification of the group is determined from the agents’ classifications by expectation-maximization techniques, with preference to the recognized experts. The algorithm was verified and tested on simulated and real-world datasets and benchmarked against known existing algorithms. We show that such a method reduces incorrect classifications and effectively solves the problem of unsupervised collaborative classification under uncertainty, while outperforming other known methods.

Due to the COVID-19 crisis, the education sector has been shifted to a virtual environment. Monitoring the engagement level and providing regular feedback during e-classes is one of the major concerns, as this facility lacks in the e-learning environment due to no physical observation of the teacher. According to present study, an engagement detection system to ensure that the students get immediate feedback during e-Learning. Our proposed engagement system analyses the student’s behaviour throughout the e-Learning session. The proposed novel approach evaluates three modalities based on the student’s behaviour, such as facial expression, eye blink count, and head movement, from the live video streams to predict student engagement in e-learning. The proposed system is implemented based on deep-learning approaches such as VGG-19 and ResNet-50 for facial emotion recognition and the facial landmark approach for eye-blinking and head movement detection. The results from different modalities (for which the algorithms are proposed) are combined to determine the EI (engagement index). Based on EI value, an engaged or disengaged state is predicted. The present study suggests that the proposed facial cues-based multimodal system accurately determines student engagement in real time. The experimental research achieved an accuracy of 92.58% and showed that the proposed engagement detection approach significantly outperforms the existing approaches.

An essential component of smart grid applications is the ability to solve the power flow (PF) problem in real-time. As numerical methods are too slow, the use of neural networks (NNs) is rapidly increasing. Graph Neural Networks (GNNs) and their variants have become one of the leading methods to learn graph representations. Power systems and in particular, distribution systems can be represented as graphs, and are characterized by often topology changes, which makes the consideration of the topology structure to be an important aspect when searching for a solution approach. Although GNNs have promising results for certain applications such as computer vision ones, considering its limitations, it still has a long way to go until becoming a leading candidate for PF based applications. This paper highlights the existing gaps and challenges in fully accepting ANNs and particularly GNNs as real-time solution engines for the PF problem in DSs. These gaps are analyzed under three categories: suitable architectures for the solution of the PF problem in DS, explicit vs. implicit incorporation of the DS topology information impact on the models’ generalization, and the limiting factors for GNNs implementation aimed at the solution of the PF problem in DSs. The paper also includes a discussion, suggestions and insights of overcoming these gaps in future research.

In this article, we utilize an explainable AI approach, the Explainable Boosting Machine (EBM), to perform feature analysis on an extensive Internet of Things (IoT) dataset collected from real-world devices. After feature selection and data processing, our training dataset includes 2.88 million traffic data instances, categorized into six classes (DDoS, DoS, Mirai, Recon, Spoofing, and Benign). The EBM trained on this dataset achieved a impressive accuracy rate of 99.4% and an F1 score of 92.8%. Using the resultant model, we interpreted its predictions based on feature importance. The identified feature importance aligned well with established cybersecurity principles, indicating the model's potential. However, our analysis revealed that the machine learning model's predictions were strongly tied to the specific characteristics of the training IoT dataset, thereby raising concerns about the model's reliability when applied to real-world attack detection. Future research could explore the use of more diverse and balanced datasets or the applicability of the machine learning model in different IoT contexts, aiming to enhance the model's generalizability and practical relevance.

Over the period of time, weather conditions have changed drastically. These changes have led to many serious issues such as soil nutrition depletion, reduction in soil water holding capacity, and increase in pesticide resistant weeds. Precision Agriculture (PA) is the pioneer that is changing the way people farm, offering benefits in profitability, productivity, sustainability, crop quality, environmental protection, on-farm quality of life, food safety, and rural economic development. In addition, with the aid of the recent advancements in the sensor technology, it can develop into an intelligent crop production system. Specifically, remote sensing technology that allows non-destructive acquisition of information about the earth’s surface can facilitate the implementation of PA. In this paper, a specific area of Punjab is analyzed for crop production using remote sensing data. To increase crop production and make management decisions, PA entails the application of remote sensing technology.

The outbreak of COVID-19 has caused an unprecedented increase in the usage of e-Learning platforms. The closure of educational institutions globally has significantly impacted the traditional education system. Without physical interaction, engaging students in an e-Learning environment has become a major challenge for teachers and e-Learning platform providers. Unlike a face-to-face classroom, where teachers can easily monitor students’ behaviour and adapt the learning content according to their needs, this is impossible in an e-Learning environment. This study introduces an optimized deep learning-based approach for detecting student engagement levels to ensure students remain connected to the learning process. The proposed optimized model uses facial emotion recognition and head movement detection to track real-time engagement levels for big data as real-time-based face and head datasets are analysed. The system leverages facial landmark detection to monitor head movements and deep learning models such as VGG19 and ResNet50 for facial emotion recognition. The system combines the output of both approaches to accurately predict the student engagement state as either ‘engaged’ or ‘disengaged’ with an accuracy rate of 91.67%.

Ensuring student engagement is crucial for effective learning outcomes in any classroom setting, including e-learning environments. However, the absence of immediate supervision in online classes makes monitoring and maintaining student attentiveness difficult. To address this challenge, this study proposes a cognitive state detection system that continuously monitors the facial emotion of the learner in an adaptive learning environment. The algorithm is proposed to detect cognitive states such as attentiveness and inattentiveness. The system has been implemented on four separate databases and evaluated using three ensemble models: FT-EDFA, FC-EDFA, and OT-EDFA. The ensemble models have been created by applying transfer learning to two popular pre-trained models, VGG19 and ResNet50, which can learn useful features from facial images for emotion recognition tasks. Combining the features learned by both models, the ensemble approach can achieve better performance in recognising facial emotions. The proposed system can provide continuous feedback to instructors, enabling them to adjust their teaching methods to maintain student engagement and interest. The study has achieved promising results, surpassing the performance of existing methods with recognition rates of 93.11%, 92.34%, and 91.12% on the newly created dataset. By detecting cognitive states in online learners, the proposed system can help instructors understand how engaged and interested their students are during classes. Overall, facial emotion recognition can be useful for improving the quality of e-learning platforms and enhancing student learning outcomes.