International Journal of Intelligent Systems最新文献

The emergence of large language models (LLMs) has substantially changed the artificial intelligence field, enabling its wide use over different domains. As various LLM alternatives have been developed, the current study proposes a novel decision-support framework for evaluating and benchmarking LLMs based on multicriteria decision-making (MCDM) techniques. In the proposed framework, an improved version of the best-worst method (BWM) is proposed to effectively reduce the computational complexity of assigning a critical weight for the evaluation criteria of LLMs. Then, the improved BWM is integrated with the combined compromise solution (CoCoSo) method for ranking LLM alternatives. Findings show that the improved BWM successfully computes the criteria weights with low computational complexity compared to the original BWM. According to the enhanced BWM, the ‘factual errors’ criterion received the highest significant weight (0.2681), while the ‘logical inconsistencies’ criteria obtained the lowest (0.0827). The rest of the criteria were distributed in between that range. Subsequently, CoCoSo ranked the involved LLM alternatives in two different runs based on the extracted weights. Sensitivity analysis was employed to evaluate the effect of the assessment criteria on LLMs’ evaluation.

Chronic obstructive pulmonary disease (COPD) is a complex and heterogeneous disease, which presents a significant challenge in identifying patients at high risk of short- and medium-term mortality. Such complexity poses challenges to clinical decision-making and the effective planning of end-of-life care in these patients. This study proposes the development of a novel intelligent clinical decision support system, designed to predict 1-year mortality in COPD patients following an acute exacerbation. The system is constructed upon a database of over 500 patients, comprising demographic, clinical, and social variables. First, a feature selection process is conducted to identify the variables that possess the greatest predictive power. Based on these, the data for each patient are encapsulated in a pseudosymbol construct that represents and consolidates them. The construction of the pseudosymbol comprises two distinct steps: (1) transforming the variables into a sound composition and (2) generating the corresponding spectrogram, which constitutes a visual representation (i.e., an image). The system employs a convolutional neural network, SqueezeNet, as the inference engine to calculate the 1-year mortality risk based on the images. Ten percent of the data was reserved for testing the system, achieving an area under the ROC curve (AUC) close to 0.85, indicating a high predictive power. Despite this promising initial result, further clinical validations in real-world settings will be necessary to confirm the system’s applicability and usefulness.

The increasing complexity of video tampering techniques poses a significant threat to the integrity and security of Internet of Multimedia Things (IoMT) ecosystems, particularly in resource-constrained edge-cloud infrastructures. This paper introduces Multiscale Gated Multihead Attention Depthwise Separable CNN (MGMA-DSCNN), an advanced deep learning framework specifically optimized for real-time tampered video detection in IoMT environments. By integrating lightweight convolutional neural networks (CNNs) with multihead attention mechanisms, MGMA-DSCNN significantly enhances feature extraction while maintaining computational efficiency. Unlike conventional methods, this approach employs a multiscale attention mechanism to refine feature representations, effectively identifying deepfake manipulations, frame insertions, splicing, and adversarial forgeries across diverse multimedia streams. Extensive experiments on multiple forensic video datasets—including the HTVD dataset—demonstrate that MGMA-DSCNN outperforms state-of-the-art architectures such as VGGNet-16, ResNet, and DenseNet, achieving an unprecedented detection accuracy of 98.1%. Furthermore, by leveraging edge-cloud synergy, our framework optimally distributes computational loads, effectively reducing latency and energy consumption, making it highly suitable for real-time security surveillance and forensic investigations. These advancements position MGMA-DSCNN as a scalable, high-performance solution for next-generation intelligent video authentication, offering robust, low-latency detection capabilities in dynamic and resource-constrained IoMT environments.

Neural network–based models have shown tremendous achievements in various fields. However, standard AI-based systems suffer from catastrophic forgetting when undertaking sequential learning of multiple tasks in dynamic environments. Continual learning has emerged as a promising approach to address catastrophic forgetting. It enables AI systems to learn, transfer, augment, fine-tune, and reuse knowledge for future tasks. The techniques used to achieve continual learning are inspired by the learning processes of the human brain. In this study, we present a comprehensive review of research and recent developments in continual learning, highlighting key contributions and challenges. We discuss essential functions of the biological brain that are pivotal for achieving continual learning and map these functions to the recent machine-learning methods to aid understanding. Additionally, we offer a critical review of five recent types of continual learning methods inspired by the biological brain. We also provide empirical results, analysis, challenges, and future directions. We hope that this study will benefit both general readers and the research community by offering a complete picture of the latest developments in this field.

In the area of ecology, counting animals to estimate population size and types of species is important for the wildlife conservation. This includes analysing massive volumes of image, video or audio/acoustic data and traditional counting techniques. Automating the process of identifying, classifying and counting animals would be helpful to researchers as it will phase out the tedious human–labour tasks of manual counting and labelling. The intention of this work is to address manual identification and counting methods of images by implementing an automated solution using computer vision and deep learning. This study applies a classification model to classify species and trains an object detection model using deep convolutional neural networks to automatically identify and determine the count of four mammal species in 3304 images extracted from camera traps. The image classification model reports a classification accuracy of 98%, and the YOLOv8 object detection model automatically detects buffalo, elephant, rhino and zebra school mean average precision of 50 of 89% and mean average precision of 50–95 of 72.2% and provides an accurate count over all animal classes. Furthermore, it performs well across various image scenarios such as blurriness, day, night and images displaying multiple species compared to the RT-DETR model. The results of the study display that the application of computer vision and deep learning methods on data-scarce and data-enriched scenarios, respectively, can conserve biologists and ecologists an enormous amount of time used on time-consuming human tasks methods of analysis and counting. The high-performing deep learning models developed capable of accurately classifying and localising multiple species can be integrated into the existing conservation workflows to process large volumes of camera trap images in real time. This integration can significantly reduce the manual labour required for labelling and counting, improve the consistency and speed of wildlife surveys and enable timely decision-making in habitat protection, population assessment and antipoaching initiatives. Additionally, these automated identification techniques can contribute towards enhancing wildlife conservation and future studies.

Unmanned aerial vehicles (UAVs) have been employed for a variety of inspection and monitoring tasks, including agricultural applications and search and rescue (SAR) in remote areas. However, traditional monitoring methods tend to focus on optimizing one aspect. This study aims to propose a complete framework by integrating advanced methods to provide a robust and accurate path coverage solution. The combination of edge detection and area decomposition with a pathfinding algorithm can improve the overall performance. An effective edge detection model is developed that simultaneously detects the boundary and segments the area of interest (AOI) from the aerial land images and provides precise area mapping of the area. An intuitive grid decomposition with grid-to-graph mapping improves the flexibility of the area decomposition and ensures maximal coverage and safe operation routes for the UAVs. Finally, a robust modified simulated annealing (MSA) algorithm is introduced to determine the shortest path coverage route. The performance of the proposed methodology is tested on aerial imagery. Area decomposition ensures that there are no gaps in the AOI during the coverage planning. The MSA algorithm obtains the minimum length cost, charge consumption cost, and minimum number of turns to cover the area. It is shown that the integration of these techniques enhances the performance of the coverage path planning (CPP). A comparison of the proposed approach with benchmark algorithms further demonstrates its effectiveness. This study contributes to creating a complete CPP application for UAVs, which may assist with precision agriculture as well as safe and secure rescue operations.