Original/Review Paper
H.5. Image Processing and Computer Vision
Khosro Rezaee
Abstract
Bipolar disorder (BD) remains a pervasive mental health challenge, demanding innovative diagnostic approaches beyond traditional, subjective assessments. This study pioneers a non-invasive handwriting-based diagnostic framework, leveraging the unique interplay between psychological states and motor expressions ...
Read More
Bipolar disorder (BD) remains a pervasive mental health challenge, demanding innovative diagnostic approaches beyond traditional, subjective assessments. This study pioneers a non-invasive handwriting-based diagnostic framework, leveraging the unique interplay between psychological states and motor expressions in writing. Our hybrid deep learning model, combining ResNet for intricate feature extraction and external attention mechanisms for global pattern analysis, achieves a remarkably high accuracy 99%, validated through Leave-One-Subject-Out (LOSO) cross-validation. Augmented with advanced data preprocessing and augmentation techniques, the framework adeptly addresses dataset imbalances and handwriting variability. For the first time, Persian handwriting serves as a medium, bridging cultural gaps in BD diagnostics. This work not only establishes handwriting as a transformative tool for mental health diagnostics but also sets the stage for accessible, scalable, and culturally adaptive solutions in global mental healthcare.
Original/Review Paper
B.3. Communication/Networking and Information Technology
Ali Abdi Seyedkolaei
Abstract
Deploying multiple sinks instead of a single sink is one possible solution to improve the lifetime and durability of wireless sensor networks. Using multiple sinks leads to the definition of a problem known as the sink placement problem. In this context, the goal is to determine the optimal locations ...
Read More
Deploying multiple sinks instead of a single sink is one possible solution to improve the lifetime and durability of wireless sensor networks. Using multiple sinks leads to the definition of a problem known as the sink placement problem. In this context, the goal is to determine the optimal locations and number of sink nodes in the network to maximize the network's lifetime. In this paper, we propose a dynamic sensor assignment algorithm to address the sink placement problem and evaluate its performance against existing solution methods on a diverse set of instances. We conducted experiments in two stages. In the first stage, based on random instances and compared to the exact computational method using the CPLEX solver, and in the second stage, based on real-world instances compared to MC-JMSP (Model-Based Clustering- Joint Multiple Sink Placement) method. The results obtained in the first stage of the experiments indicate the superiority of the dynamic sensor assignment algorithm in runtime for all instances. Furthermore, the solution obtained by the dynamic sensor assignment algorithm is very close to the solution obtained by the CPLEX solver. In particular, the percentage error of the solution found by the proposed method compared to CPLEX in all experimented instances is less than 0.15%, indicating the effectiveness of the proposed method in finding the appropriate solution for assigning sensors to sinks. Also, the results of the second stage experiments show the superiority of the proposed method in both execution time and energy efficiency compared to the MC-JMSP method.
Original/Review Paper
H.3. Artificial Intelligence
Habib Khodadadi; Vali Derhami
Abstract
The exploration-exploitation trade-off poses a significant challenge in reinforcement learning. For this reason, action selection methods such as ε-greedy and Soft-Max approaches are used instead of the greedy method. These methods use random numbers to select an action that balances exploration ...
Read More
The exploration-exploitation trade-off poses a significant challenge in reinforcement learning. For this reason, action selection methods such as ε-greedy and Soft-Max approaches are used instead of the greedy method. These methods use random numbers to select an action that balances exploration and exploitation. Chaos is commonly utilized across various scientific disciplines because of its features, including non-periodicity, unpredictability, ergodicity and pseudorandom behavior. In this paper, we employ numbers generated by different chaotic systems to select action and identify better maps in diverse states and quantities of actions. Based on our experiments on various environments such as the Multi-Armed Bandit (MAB), taxi-domain, and cliff-walking, we found that many of the chaotic methods increase the speed of learning and achieve higher rewards.
Methodologies
H.6.3.2. Feature evaluation and selection
Sayyed Mohammad Hoseini; Majid Ebtia; Mohanna Dehgardi
Abstract
The abundance of high dimensional datasets and the computational limitations of data analysis processes in applying to high-dimensional data have made clear the importance of developing feature selection methods. The negative impact of irrelevant variables on prediction and increasing unnecessary calculations ...
Read More
The abundance of high dimensional datasets and the computational limitations of data analysis processes in applying to high-dimensional data have made clear the importance of developing feature selection methods. The negative impact of irrelevant variables on prediction and increasing unnecessary calculations due to the redundant attributes lead to poor results or performance of the classifiers. Feature selection is, therefore, applied to facilitate a better understanding of the datasets, reduce computational time, and enhance prediction accuracy. In this research, we develop a composite method for feature selection that combines support vector machines and principal component analysis. Then the method is implemented to the -nearest neighbor and the Naïve Bayes algorithms. The datasets utilized in this study consist of three from the UCI Machine Learning Repository, used to assess the performance of the proposed models. Additionally, a dataset gathered from the central library of Ayatollah Boroujerdi University was considered. This dataset encompasses 1,910 instances with 30 attributes, including gender, native status, entry term, faculty code, cumulative GPA, and the number of books borrowed. After applying the proposed feature selection method, an accuracy of 70% was obtained with only five features. Experimental results demonstrate that the proposed feature selection method chooses appropriate feature subset. The approach yields enhanced classification performance, as evaluated by metrics such as accuracy, -score and Matthews correlation coefficient.
Applied Article
I.3.7. Engineering
Saeed Khosroabadi; Hussein Aad Alaboodi
Abstract
In the context of advancing sixth-generation (6G) communication networks, ensuring high-quality user coverage across varying geographic landscapes remains a paramount objective. Terrestrial base stations conventionally provide this coverage; however, they are susceptible to disruption due to adverse ...
Read More
In the context of advancing sixth-generation (6G) communication networks, ensuring high-quality user coverage across varying geographic landscapes remains a paramount objective. Terrestrial base stations conventionally provide this coverage; however, they are susceptible to disruption due to adverse environmental conditions. Consequently, the integration of airborne mobile stations is pivotal for continued user coverage support. Among the viable solutions for terrestrial station augmentation, the deployment of drone base stations (DBS) emerges as the optimal substitute. Nonetheless, the establishment of a drone-based infrastructure presents challenges in terms of time and cost efficiency. Thus, the strategic positioning of DBSs, aimed at maximizing user coverage while simultaneously minimizing path loss and the number of drones required, is essential to achieving efficient and high-quality service provisioning. This study introduces a novel and optimized DBS placement strategy utilizing the Marine Predators Algorithm (MPA)—a recent metaheuristic renowned for its potent resistance to entrapment in local optima. Through simulation, we demonstrate that our proposed methodology distinctly surpasses analogous approaches with regards to optimization of path loss and user coverage. Simulation outcomes reveal average path losses of 71.75 dB for the Gray Wolf Optimization (GWO), 75.78 dB for the Weighted Time-Based Non-Orthogonal Multiple Access (TW-NOMA), and a significantly reduced 56.13 dB for our proposed MPA-based method, thereby indicating a substantial decrease of at least 15 dB in path loss compared to current techniques.
Original/Review Paper
F.2.7. Optimization
Seyed Morteza Babamir; Narges Zahiri
Abstract
Web service composition represents a graph of interacting services designed to fulfill user requirements, where each node denotes a service, and each edge represents an interaction between two services. A few candidates with different quality attributes exist on the web for conducting each web service. ...
Read More
Web service composition represents a graph of interacting services designed to fulfill user requirements, where each node denotes a service, and each edge represents an interaction between two services. A few candidates with different quality attributes exist on the web for conducting each web service. Consequently, numerous compositions with identical functionality but differing quality attributes can be formed, making the near-optimal composition selection an NP-hard problem. This paper proposes a tool-supported Evolutionary Optimization Algorithm (EOA) for near-optimal composition selection. The proposed EOA is a Discretized and Extended Gray Wolf Optimization (DEGWO) algorithm. This approach first discretizes the continuous solution space and then extends the functionality of GWO to identify global near-optimal solutions while accelerating solution convergence. DEGWO was evaluated in comparison with other related methods in terms of metrics. Experimental results showed DEGWO achieved average improvements of 8%, 39%, and 5% in terms of availability, 36%, 43%, and 30% in terms of response time, and 65%, 53%, and 51% in terms of cost compared to the three leading algorithms, RDGWO+GA, HGWO, and SFLAGA, respectively.
Original/Review Paper
F.1. General
Farzad Zandi; Parvaneh Mansouri; Reza Sheibani
Abstract
In the field of optimization, metaheuristic algorithms have garnered significant interest. These algorithms, which draw inspiration from natural selection, evolution, and problem-solving strategies, offer an alternative approach to solving complex optimization problems. Unlike conventional software engineering ...
Read More
In the field of optimization, metaheuristic algorithms have garnered significant interest. These algorithms, which draw inspiration from natural selection, evolution, and problem-solving strategies, offer an alternative approach to solving complex optimization problems. Unlike conventional software engineering methods, metaheuristics do not rely on derivative calculations in the search space. Instead, they explore solutions by iteratively refining and adapting their search process. The no-free-lunch (NFL) theorem proves that an optimization scheme cannot perform well in dealing with all optimization challenges. Over the last two decades, a plethora of metaheuristic algorithms has emerged, each with its unique characteristics and limitations. In this paper, we propose a novel meta-heuristic algorithm called ISUD (Individuals with Substance Use Disorder) to solving optimization problems by examining the clinical behaviors of individuals compelled to use drugs. We evaluate the effectiveness of ISUD by comparing it with several well-known heuristic algorithms across 44 benchmark functions of varying dimensions. Our results demonstrate that ISUD outperforms these existing methods, providing superior solutions for optimization problems.
Original/Review Paper
H.3. Artificial Intelligence
Thomas Njoroge Kinyanjui; Kelvin Mugoye; Rachael Kibuku
Abstract
This paper presents a Multi-Head Self-Attention Fusion Network (MHSA-FN) for real-time crop disease classification, addressing key limitations in existing models, including suboptimal feature extraction, inefficient feature recalibration, and weak multi-scale fusion. Unlike prior works that rely solely ...
Read More
This paper presents a Multi-Head Self-Attention Fusion Network (MHSA-FN) for real-time crop disease classification, addressing key limitations in existing models, including suboptimal feature extraction, inefficient feature recalibration, and weak multi-scale fusion. Unlike prior works that rely solely on CNNs or transformers, MHSA-FN integrates MobileNetV2, EfficientNetV2, and Vision Transformers (ViTs) with a structured multi-level attention framework for enhanced feature learning. A gated fusion mechanism and a Multiscale Fusion Module (MSFM) optimize local texture details and global spatial relationships. The model was trained on a combined dataset of PlantVillage and locally collected images, improving adaptability to real-world conditions. It achieved 98.66% training accuracy and 99.0% test accuracy across 76 disease classes, with 99.34% precision, 99.01% recall, and 99.04% F1 score. McNemar’s test (p = 0.125) and Bayesian superiority probability (0.851) validated its robustness. Confidence variance analysis (0.000010) outperformed existing models, demonstrating MHSA-FN as a scalable, high-performance AI solution for precision agriculture in resource-constrained environments.
Original/Review Paper
I.5. Social and Behavioral Sciences
Havva Alizadeh Noughabi
Abstract
Social media platforms have transformed information consumption, offering personalized features that enhance engagement and streamline content discovery. Among these, the Twitter Lists functionality allows users to curate content by grouping accounts based on shared themes, fostering focused interactions ...
Read More
Social media platforms have transformed information consumption, offering personalized features that enhance engagement and streamline content discovery. Among these, the Twitter Lists functionality allows users to curate content by grouping accounts based on shared themes, fostering focused interactions and diverse perspectives. Despite their widespread use, the relationship between user-generated content and List subscription behaviors remains insufficiently explored. This study examines the alignment between users' post topics and their subscribed Lists, along with the influence of activity levels on this alignment. The role of content diversity in shaping subscription patterns to Lists covering a range of topics is also investigated. Additionally, the extent to which the affective characteristics—sentiment and emotion—of user posts correspond with the emotional tone of subscribed List content is analyzed. Utilizing a comprehensive Twitter dataset, advanced techniques for topic modeling, sentiment analysis, and emotion extraction were applied, and profiles for both users and Lists were developed to facilitate the exploration of their interrelationship. These insights advance the understanding of user interactions with Lists, informing the development of personalized recommendation systems and improved content curation strategies, with broad implications for social media research and platform functionality.
Original/Review Paper
H.3.8. Natural Language Processing
Milad Allahgholi; Hossein Rahmani; Parinaz Soltanzadeh
Abstract
Stance detection is the process of identifying and classifying an author's point of view or stance towards a specific target in a given text. Most of previous studies on stance detection neglect the contextual information hidden in the input data and as a result lead to less accurate results. In this ...
Read More
Stance detection is the process of identifying and classifying an author's point of view or stance towards a specific target in a given text. Most of previous studies on stance detection neglect the contextual information hidden in the input data and as a result lead to less accurate results. In this paper, we propose a novel method called ConSPro, which uses decoder-only transformers to consider contextual input data in the process of stance detection. First, ConSPro applies zero-shot prompting of decoder only transformers to extract the context of target in the input data. Second, in addition to target and input text, ConSPro uses the extracted context as the third type of parameter for the ensemble method. We evaluate ConSPro on SemEval2016 and the empirical results indicate that ConSPro outperforms the non-contextual approaches methods, on average 9% with respect to f-measure. The findings of this study show the strong capabilities of zero-shot prompting for extracting the informative contextual information with significantly less effort comparing to previous methods on context extraction.