In environments where human presence is restricted due to safety concerns, cognitive robots can play a vital role in executing tasks. While robotic systems have made impressive advances in tasks such as object recognition, they still fall short in terms of true spatial understanding. Robots are not yet able to understand the 3D spatial semantics and contextual meaning of their surroundings for navigation, nor can they effectively handle objects in complex tasks. This research tackles this obstacle by developing a computational agent capable of learning cognitive maps from spatial data inputs in a simulated environment, mimicking the functionalities of grid and place neurons. To emulate the grid neuron's ability to generate periodic hexagonal grid-like patterns from body movements in 3-dimensional space, a novel Octant model is introduced. Additionally, a place-grid neuron interaction system is proposed to predict environmental sensations from body movements, facilitating cognitive map formation and learning mechanisms. The model was experimentally tested through a set of simulation-based experiments including: (i) a grid-based spatial arena intended to measure the accuracy of memory retrieval and encoding, (ii) a morphologically perturbed environment to measure resiliency to deformations in the trajectories, and (iii) a collection of more (object) recognition activities, to test resiliency to distortions in the object identification task. The quantitative data indicate steady high recall confidence and cosine similarity during the navigational trials, and a controlled growth of the place-neuron memory capacity and a strong stabilization of spatial representations in different environments. In obstacle-rich conditions, the model achieves a mean Absolute Trajectory Error (ATE) of 0.42 m and a root-mean-square error (RMSE) of 0.46 m, confirming that localization error remains limited even without an explicit mapping structure. These results provide a substantive confirmation of the fact that the interference-based grid place representation allows reliable spatial localization, mnemonic recall and navigational performance in both limited and irregular operational conditions.

Stock market forecasting is not an easy task to undertake because of the volatility and the price movements which are non-stationary. In this work, the authors propose a hybrid deep learning architecture that combines the use of Discrete Wavelet Transform (DWT)-based feature extraction with multi-architecture aggregation with LSTM, GRU, RNN, and CNN models. The model was tested on six Indian stocks based on ten-years of daily historical returns under a rolling walk-forward validation procedure, with the model being trained on a five-year window and tested on out-of-sample periods. It has been shown experimentally that the hybrid aggregation method gives smaller prediction errors than individual architectures. Using the proposed model on the NIFTY 50 index, the RMSE was 0.2102, and the directional accuracy was 65.18%, indicating improved predictive stability compared to baseline models. An ablation analysis also supports the fact that wavelet-based pre-processing helps to reduce errors and increase consistency of trends. These results indicate that wavelet-based feature extraction with heterogeneous deep learning models can be more robust when used in daily stock forecasting. In future work, statistical significance tests and cost-based trading simulations may be incorporated to further assess practical relevance.

In real-world clinical settings, the growing number of patients and the shortage of experienced ophthalmologists make early and accurate diagnosis of retinal diseases increasingly challenging. Cataracts, diabetic retinopathy, and glaucoma are some of the most common causes of lifelong blindness around the world. This is why there is a need for automated diagnostic systems that can accurately diagnose and interpret clinical data. The major goal of this work is to find out if a deep learning architecture based on EfficientNetB3 and Explainable Artificial Intelligence (XAI) can accurately classify multiple types of retinal diseases while still being clear to doctors. The proposed system categorizes retinal fundus images into four groups: cataract, diabetic retinopathy, glaucoma, and normal. The dataset consisted of a balanced and publicly accessible collection of 4,217 retinal fundus pictures, processed using standard preprocessing techniques to enhance their generalizability. We chose EfficientNetB3 as the main architecture since it is better at extracting features, and we compared it to the standard convolutional neural network baselines to show how useful it is. The proposed model achieved an accuracy of 97%, outperforming ResNet50 (91%) and VGG16 (87%). The high precision, recall, and F1-scores (0.94 – 1.00), the Cohen’s kappa of 0.95, and the low logarithmic loss of 0.10 all point to reliable predictions. The receiver operating characteristic (ROC) analysis yielded an AUC of 1.00 across all disease classes. Gradient-weighted Class Activation Mapping (Grad-CAM) was utilized to address the interpretability deficit in deep learning-based medical systems and to pinpoint clinically significant retinal regions that influence model predictions. The results indicate that employing XAI alongside EfficientNetB3 enhances both diagnostic precision and interpretability, hence validating its suitability as a transparent decision-support system for the automated screening of retinal disorders. 

Reducing student anxiety is a critical factor in enhancing academic performance, with timely identification of mental health concerns serving as a key component of effective educational support strategies. This study presents the development and integration of a classification model designed to detect students’ current programming anxiety levels based on academic, demographic, and behavioral attributes. The model was trained and evaluated using student data collected from students enrolled in computing-related programs, and its performance was benchmarked against Support Vector Machine, Logistic Regression, Random Forest, J48 Decision Tree, and Naive Bayes classifiers. The Support Vector Machine achieved excellent performance, with an F-measure of 95.97%, an accuracy of 97.11%, precision of 94.20%, recall of 97.86%, and a Cohen’s kappa of 0.937, indicating strong agreement between predicted and actual classifications. Error-based metrics further supported the model’s reliability, with a Mean Absolute Error (MAE) of 0.0289, Root Mean Squared Error (RMSE) of 0.1641, Relative Absolute Error (RAE) of 0.04%, and Root Relative Squared Error (RRSE) of 34.40%. To enhance practical utility, the model was integrated into a web-based student information system that generates real-time classifications and visualizations, supporting educators in recognizing students who may require additional support. While the model provides valuable insights for identifying students exhibiting current programming anxiety, future work incorporating longitudinal data is needed to enable true predictive capabilities for early intervention.

Minimizing energy consumption and carbon emissions while maintaining system performance is a critical challenge in cloud task scheduling. This paper presents a multi-objective scheduling framework based on a Memetic Algorithm (MA) designed to optimize task-to-VM mapping with respect to energy efficiency, carbon footprint, and throughput. The algorithm employs a weighted fitness function that integrates actual and idle energy usage, simulated time-varying carbon intensity, and task throughput. To enhance solution quality, MA combines global evolutionary operations (selection, crossover, mutation) with local search heuristics that adaptively refine candidate solutions based on workload characteristics and green energy opportunities. The carbon emission model incorporates dynamic emission factors (γ) derived from location- and time-sensitive datasets, reflecting real-world variability in grid carbon intensity. The proposed method is evaluated using the NASA Ames iPSC/860 workload under both low and high resource utilization scenarios. Comparative results demonstrate that the proposed MA approach achieves a 20.2% reduction in carbon emissions, reduces energy consumption by 17.7%, and enhances throughput by 21.2% over conventional techniques such as HDDPGTS and RAPTS, while also ensuring competitive performance in terms of makespan and resource utilization. These improvements underscore the potential of memetic-based hybrid scheduling to support environmentally sustainable and performance-efficient cloud infrastructures. The findings highlight the importance of integrating eco-aware intelligence into task scheduling policies, particularly for mission-critical and energy-intensive cloud applications.

Automatic human activity recognition has many applications in smart environments, including smart homes, smart cities, smart industries, and smart healthcare centers. While performing the activities by the participants, ambient or body-worn sensors can measure physical movements and those data can be used to develop machine learning models for recognizing those activities. In this study, we have proposed a deep convolutional neural network (DCNN) based method for recognizing human activities using body-worn sensors’ time-series data after extensive data analysis. High-quality data is generated and balanced using a preprocessing chain for human activity recognition based on data analysis. The preprocessed data is segmented using a constant-size sliding window. We developed several different DCNN models using random searches and based on validation accuracy we selected the best one for further training and testing. The outputs of the selected model serve as the final predicted activities. We assessed our method on three popular and standard datasets: PAMAP2, WISDM_ar_v1.1, and UCI-HAR, and achieved 98.11%, 98.48%, and 93.25% accuracies for subject-dependent case and 90.27%, 94.51%, and 98.67% accuracies for subject-independent case. The performances of the experimental results are measured using several evaluation metrics and measures that demonstrate the strength of the proposed model over the state-of-the-art. 

This study presents the Multi-Modal Deep Fusion Network to identify cotton leaf diseases. Initially, the images are collected from the Kaggle cotton disease dataset. The dataset is preprocessed, and data augmentation is applied exclusively to the training set to prevent data leakage. The VGG-16-based Faster Region-based Convolutional Neural Network model is used for lesion detection and region of interest localization by generating bounding boxes around diseased areas. Both handcrafted features (including shape descriptors and color moments) and deep learning features are used for feature extraction. The extracted features are optimized using the Snowy Wolf Optimization algorithm which combines Snow Leopard Optimization and Grey Wolf Optimization. The proposed method achieved 98.4% accuracy, 98.6% sensitivity, and 98.8% F-score, consistently outperforming existing methods under identical experimental settings. While the proposed framework demonstrated promising performance on the evaluated dataset, further validation on larger and more diverse field datasets is required to comprehensively assess its generalization capability.

Machine learning models, particularly deep learning architectures, achieve high performance in prediction tasks but remain susceptible to adversarial attacks. This study aims to enhance the robustness of Convolutional Neural Networks (CNNs), Deep Neural Networks (DNNs), and Recurrent Neural Networks (RNNs), thereby improving the security of machine learning systems. A three-step approach is adopted. First, benign sample classification is performed using the MNIST benchmark dataset. Second, adversarial attacks, namely Projected Gradient Descent (PGD), DeepFool (DF), and the Fast Gradient Sign Method (FGSM), are launched on the trained models, resulting in significant performance degradation. Based on the biased outputs induced by adversarial perturbations, an adversarial detection model is subsequently established. Third, to counteract these attacks, various defense strategies, including adversarial training, defensive distillation, autoencoder-based denoising, ensemble methods, and feature squeezing are employed and evaluated using standard performance metrics and graphical analyses. The results indicate that, in the absence of defense mechanisms, PGD attacks lead to accuracy drops of approximately 27% in CNNs, 83% in DNNs, and 90% in RNNs, demonstrating severe model vulnerabilities. However, when defense strategies are applied, all models recover to an accuracy of at least 98.9%, with adversarial training improving performance under attack by up to 90%. Among the evaluated models, CNNs exhibit the highest baseline robustness, whereas DNNs and RNNs rely more heavily on defense mechanisms to maintain performance. These findings provide valuable insights into the development of secure and resilient machine learning systems capable of mitigating adversarial threats.

Social media’s worldwide expansion over the past two decades has significantly altered the dissemination of extremist narratives, creating both challenges and opportunities for counterterrorism efforts. Addressing critical gaps in the detection and classification of extremist content on social media platforms, this research supports earlier-stage analytical assessment for law enforcement and security agencies. Using datasets from the publicly available Global Terrorism Database (GTD, n > 209,000 incidents) and a curated corpus of labeled tweets (n = 17,410), a hybrid framework integrating machine learning and deep learning models through a late-fusion stacking architecture is developed. The proposed ensemble leverages contextual indicators derived from historical terrorism data alongside linguistic and behavioral signals from social media content to distinguish extremist from non-extremist activity. Evaluated under strict temporal validation, the model achieves an accuracy of 98.52%, precision of 97.01%, recall of 99.66%, and an AUC of 0.92 under controlled experimental conditions. To address ethical and transparency considerations, Shapley Additive exPlanations (SHAP) are employed to enhance interpretability in automated decision-making. While the reported results reflect dataset-specific evaluation, the findings demonstrate that integrating historical terrorism intelligence with temporally ordered social media analysis can support counterterrorism efforts by mitigating digital radicalization pathways and associated downstream physical security risks linked to terrorism and extremism through earlier analytical intervention.

Epilepsy is a neurological condition that affects the emotional and psychological well-being of individuals. Managing this disorder is challenging, particularly because focal seizures, which begin in specific brain regions, can sometimes evolve into generalized forms. The fundamental method for seizure identification is the analysis of the electroencephalogram (EEG), yet its manual interpretation is prone to error. In addition, the process of automated seizure detection using EEG data suffers from variations between subjects and dataset distributions as they might result in inconsistencies. Moreover, the unequal proportion of seizure to non-seizure samples increases the detection challenge. Hence, developing classification approaches capable of distinguishing and predicting the focal and generalized seizures is important for effective treatment planning. Therefore, an efficient deep learning-based focal and generalized epilepsy classification is designed in this research by considering the multimodal data. Initially, essential signals used for the validation are sourced from publicly available EEG datasets and they are converted into Short-Time Fourier Transform (STFT) images, which are designated as Feature Set 1. Next, the Sensor data used for the validation are gathered from Kaggle (https://www.kaggle.com/datasets/datasetengineer/epilepsy-dataset) and it is designated as Feature Set 2. Next, the two acquired sets of features are input to the Multilevel Spatio-Temporal Attention Fusion Network (MSTAFN) to execute the feature fusion process. Once the feature fusion procedure is completed, the fused features are given as the input to the focal and generalized epilepsy classification phase. In this phase, Adaptive Variational Autoencoders with Dense Bidirectional Gated Recurrent Units (AVDBiGRU) are employed to perform the classification process. Moreover, the focal and generalized epilepsy classification process is improved by optimizing the hyperparameters of AVDBiGRU through Fitness-based Football Optimization Algorithm (FFbOA). Finally, the focal and generalized epilepsy classified outcome is obtained from AVDBiGRU. Further, various experiments are carried out in the developed focal and generalized epilepsy classification model over the widely adopted deep learning architectures like LSTM, DCNN, InceptionV3 and BiGRU to verify their efficiency over different classes. The proposed model is evaluated on an EEG dataset containing the high-frequency oscillation (HFO) annotations from 30 pediatric patients with epilepsy and model performance is assessed by using the standard evaluation metrics that includes accuracy, sensitivity, specificity, and F1-score. The proposed model achieved 95.54% accuracy, 96.66% specificity, 93.32% F1-score and 88.53 AUC.

Population growth and pandemics like COVID-19 have led to the depletion of natural resources and an increase in hospital waste generation. This issue is particularly pressing in developing countries, where innovative solutions are needed to address the environmental and health risks associated with improper waste disposal. Traditional waste sorting methods, which rely on human intervention, are time-consuming and pose a significant risk of infection. Moreover, different categories of hospital waste require specific treatment methods. This study proposes an artificial intelligence-based approach for classifying and sorting hospital waste using Convolutional Neural Networks (CNNs). The proposed CNN model effectively identifies and categorizes various types of hospital waste, providing a sustainable solution that enhances regulatory compliance. The model leverages K-fold cross-validation and data augmentation on a publicly available, modest-sized dataset to handle variability in waste categories. The model achieved a peak classification accuracy of 97.08%, along with high precision, recall, and F1-scores, despite class imbalances and the presence of visually similar waste categories. These results highlight the potential of the model to improve hospital waste management practices, thereby reducing environmental impact and health risks.

The high volume of information from online media in Indonesia poses challenges for manual analysis in identifying emerging themes and sentiments. News headlines, as the primary element seen by the public, play a crucial role in shaping opinion; however, their massive volume and diverse themes necessitate automated approaches to identify topics and underlying sentiments. To address this challenge, this study analyzed 30,329 news headlines from the online news portal detik.com for the entire year 2024. A quantitative Natural Language Processing (NLP) framework was applied, comprising data collection via web scraping, text preprocessing, transformer-based topic modeling using BERTopic, sentiment classification using IndoBERT, and a topic-sentiment intersection analysis. Preprocessing included case folding, text cleaning, normalization of informal words, and tokenization. For lexicon-based labeling, stopword removal and stemming were applied. In contrast, transformer-based models utilized text that underwent only case folding, cleaning, and normalization to preserve contextual information. Topic modeling was performed using BERTopic, while sentiment classification (positive, negative, and neutral) used the IndoBERT model. The main objective of this study was to evaluate the combined performance of the two models in mapping dominant issues and the sentiments contained in media reports. The results showed that BERTopic successfully identified 366 topics. An evaluation of the 10 most dominant topics yielded a coherence score of 0.5145, indicating a relevant topic clustering. IndoBERT demonstrated high agreement with lexicon-generated sentiment labels, with an accuracy of 94.78%, a precision of 95.04%, a recall of 94.79%, and an F1-score of 94.81%. These findings confirm that the combination of transformer-based models is effective for in-depth discourse analysis of political news headlines from detik.com, a major Indonesian online news portal.