AI, Volume 6, Issue 1 (January 2025) – 18 articles

The relationship between the cognitive and affective dimensions of understanding has remained unexplored due to the lack of reliable methods for measuring emotions and feelings during learning. Focusing on five phases of understanding—nascent understanding, misunderstanding, confusion, emergent understanding, and deep understanding—this study introduces an AI-driven solution to measure subjective understanding by analyzing physiological activity manifested in facial expressions. To investigate these phases, 103 participants remotely worked on 15 riddles while their facial expressions were video recorded. Action units (AUs) for each phase instance were measured using AFFDEX software. AU patterns associated with each phase were then identified through the application of six supervised machine learning algorithms. Distinct AU patterns were found for all five phases, with gradient boosting machine and random forest models achieving the highest predictive accuracy. These findings suggest that physiological activity can be leveraged to reliably measure understanding. Further, they advance a novel approach for measuring and fostering understanding in educational settings, as well as developing adaptive learning technologies and personalized educational interventions. Future studies should explore how physiological signatures of understanding phases both reflect and influence their associated cognitive processes, as well as the generalizability of this study’s findings across diverse populations and learning contexts (A suite of AI tools was employed in the development of this paper: (1) ChatGPT4o (for writing clarity and reference checking), (2) Grammarly (for grammar and editorial corrections), and (3) ResearchRabbit (reference management)). Full article

Gesture recognition has a crucial role in Human–Computer Interaction (HCI) and in assisting the elderly to perform automatically their everyday activities. In this paper, three methods for gesture recognition and computer vision were implemented and tested in order to investigate the most suitable one. All methods, machine learning using IMU, machine learning on device, and were combined with certain activities that were determined during a needs analysis research. The same volunteers took part in the pilot testing of the proposed methods. The results highlight the strengths and weaknesses of each approach, revealing that while some methods excel in specific scenarios, the integrated solution of MoveNet and CNN provides a robust framework for real-time gesture recognition. Full article

Introduction: This study introduces a fully differentiable, end-to-end audio transformation network designed to overcome these limitations by operating directly on acoustic features. Methods: The proposed method employs an encoder–decoder architecture with a global conditioning mechanism. It eliminates the need for parallel utterances, intermediate phonetic representations, and speaker-independent ASR systems. The system is evaluated on tasks of voice conversion and musical style transfer using subjective and objective metrics. Results: Experimental results demonstrate the model’s efficacy, achieving competitive performance in both seen and unseen target scenarios. The proposed framework outperforms seven existing systems for audio transformation and aligns closely with state-of-the-art methods. Conclusion: This approach simplifies feature engineering, ensures vocabulary independence, and broadens the applicability of audio transformations across diverse domains, such as personalized voice assistants and musical experimentation. Full article

Background: Accurate delineation of lesions in acute ischemic stroke is important for determining the extent of tissue damage and the identification of potentially salvageable brain tissues. Automatic segmentation on CT images is challenging due to the poor contrast-to-noise ratio. Quantitative CT perfusion images improve the estimation of the perfusion deficit regions; however, they are limited by a poor signal-to-noise ratio. The study aims to investigate the potential of deep learning (DL) algorithms for the improved segmentation of ischemic lesions. Methods: This study proposes a novel DL architecture, DenseResU-NetCTPSS, for stroke segmentation using multiparametric CT perfusion images. The proposed network is benchmarked against state-of-the-art DL models. Its performance is assessed using the ISLES-2018 challenge dataset, a widely recognized dataset for stroke segmentation in CT images. The proposed network was evaluated on both training and test datasets. Results: The final optimized network takes three image sequences, namely CT, cerebral blood volume (CBV), and time to max (Tmax), as input to perform segmentation. The network achieved a dice score of 0.65 ± 0.19 and 0.45 ± 0.32 on the training and testing datasets. The model demonstrated a notable improvement over existing state-of-the-art DL models. Conclusions: The optimized model combines CT, CBV, and T_max images, enabling automatic lesion identification with reasonable accuracy and aiding radiologists in faster, more objective assessments. Full article

Background: AI-driven mental health solutions offer transformative potential for improving mental healthcare outcomes, but identifying the most effective approaches remains a challenge. This study addresses this gap by evaluating and prioritizing AI-driven mental health alternatives based on key criteria, including feasibility of implementation, cost-effectiveness, scalability, ethical compliance, user satisfaction, and impact on clinical outcomes. Methods: A fuzzy multi-criteria decision-making (MCDM) model, consisting of fuzzy TOPSIS and fuzzy ARAS, was employed to rank the alternatives, while a hybridization of the two methods was used to address discrepancies between the methods, each emphasizing distinct evaluative aspect. Results: Fuzzy TOPSIS, focusing on closeness to the ideal solution, ranked personalization of care (A5) as the top alternative with a closeness coefficient of 0.50, followed by user engagement (A2) at 0.45. Fuzzy ARAS, which evaluates cumulative performance, also ranked A5 the highest, with an overall performance rating of Si = 0.90 and utility degree Qi = 0.92. Combining both methods provided a balanced assessment, with A5 retaining its top position due to high scores in user satisfaction and clinical outcomes. Conclusions: This result underscores the importance of personalization and engagement in optimizing AI-driven mental health solutions, suggesting that tailored, user-focused approaches are pivotal for maximizing treatment success and user adherence. Full article

This study aimed to evaluate the potential of Large Language Models (LLMs) in healthcare diagnostics, specifically their ability to analyze symptom-based prompts and provide accurate diagnoses. The study focused on models including GPT-4, GPT-4o, Gemini, o1 Preview, and GPT-3.5, assessing their performance in identifying illnesses based solely on provided symptoms. Symptom-based prompts were curated from reputable medical sources to ensure validity and relevance. Each model was tested under controlled conditions to evaluate their diagnostic accuracy, precision, recall, and decision-making capabilities. Specific scenarios were designed to explore their performance in both general and high-stakes diagnostic tasks. Among the models, GPT-4 achieved the highest diagnostic accuracy, demonstrating strong alignment with medical reasoning. Gemini excelled in high-stakes scenarios requiring precise decision-making. GPT-4o and o1 Preview showed balanced performance, effectively handling real-time diagnostic tasks with a focus on both precision and recall. GPT-3.5, though less advanced, proved dependable for general diagnostic tasks. This study highlights the strengths and limitations of LLMs in healthcare diagnostics. While models such as GPT-4 and Gemini exhibit promise, challenges such as privacy compliance, ethical considerations, and the mitigation of inherent biases must be addressed. The findings suggest pathways for responsibly integrating LLMs into diagnostic processes to enhance healthcare outcomes. Full article

The growing interest in advanced large language models (LLMs) like ChatGPT has sparked debate about how best to use them in various human activities. However, a neglected issue in the debate concerning the applications of LLMs is whether they can reason logically and follow rules in novel contexts, which are critical for our understanding and applications of LLMs. To address this knowledge gap, this study investigates five LLMs (ChatGPT-4o, Claude, Gemini, Meta AI, and Mistral) using word ladder puzzles to assess their logical reasoning and rule-adherence capabilities. Our two-phase methodology involves (1) explicit instructions about word ladder puzzles and rules regarding how to solve the puzzles and then evaluate rule understanding, followed by (2) assessing LLMs’ ability to create and solve word ladder puzzles while adhering to rules. Additionally, we test their ability to implicitly recognize and avoid HIPAA privacy rule violations as an example of a real-world scenario. Our findings reveal that LLMs show a persistent lack of logical reasoning and systematically fail to follow puzzle rules. Furthermore, all LLMs except Claude prioritized task completion (text writing) over ethical considerations in the HIPAA test. Our findings expose critical flaws in LLMs’ reasoning and rule-following capabilities, raising concerns about their reliability in critical tasks requiring strict rule-following and logical reasoning. Therefore, we urge caution when integrating LLMs into critical fields and highlight the need for further research into their capabilities and limitations to ensure responsible AI development. Full article

This study investigates the evolving role of artificial intelligence (AI) in consumer decision-making, particularly in comparison to traditional human influencers. As consumer trust in social media influencers has declined, largely due to concerns about the financial motivations behind endorsements, AI tools such as ChatGPT have emerged as perceived neutral intermediaries. The research aims to understand whether AI systems can replace human influencers in shaping purchasing decisions and, if so, in which sectors. A mixed-methods approach was employed, involving a quantitative survey of 478 participants with prior experience using both AI tools and interacting with social media influencers, complemented by 15 semi-structured interviews. The results reveal that AI is favoured over human influencers in product categories where objectivity and precision are critical, such as electronics and sporting goods, while human influencers remain influential in emotionally driven sectors like fashion and beauty. These findings suggest that the future of marketing will show a reduced need for human social media influencers and may involve a hybrid model where AI systems dominate data-driven recommendations and human influencers continue to foster emotional engagement. This shift has important implications for brands as they adapt to changing consumer trust dynamics. Full article

The healthcare sector in India has experienced significant transformations owing to the advancement in technology and infrastructure. Despite these transformations, there are major challenges to address critical issues like insufficient healthcare infrastructure for the country’s huge population, limited accessibility, shortage of skilled professionals, and high-quality care. Artificial intelligence (AI)-driven solutions have the potential to lessen the stress on India’s healthcare system; however, integrating trustworthy AI in the sector remains challenging due to ethical and regulatory constraints. This study aims to critically review the current status of the development of AI systems in Indian healthcare and how well it satisfies the ethical and legal aspects of AI, as well as to identify the challenges and opportunities in adoption of trustworthy AI in the Indian healthcare sector. This study reviewed 15 articles selected from a total of 1136 articles gathered from two electronic databases, PubMed and Google Scholar, as well as project websites. This study makes use of the Preferred Reporting Items for Systematic Reviews and Meta-Analyses Extension for Scoping Reviews (PRISMA-ScR). It finds that the existing studies mostly used conventional machine learning (ML) algorithms and artificial neural networks (ANNs) for a variety of tasks, such as drug discovery, disease surveillance systems, early disease detection and diagnostic accuracy, and management of healthcare resources in India. This study identifies a gap in the adoption of trustworthy AI in Indian healthcare and various challenges associated with it. It explores opportunities for developing trustworthy AI in Indian healthcare settings, prioritizing patient safety, data privacy, and compliance with ethical and legal standards. Full article

Background: Accurately identifying the socio-demographic information of customers is crucial for utilities. It enables them to efficiently deliver personalized energy services and manage distribution networks. In recent years, machine learning-based data-driven methods have gained popularity compared to traditional survey-based approaches, owing to their time and cost efficiency, as well as the availability of a large amount of high-frequency smart meter data. Methods: In this paper, we propose a new method that harnesses the power of neural architecture search to automatically design deep neural network architectures tailored for identifying various socio-demographic information of customers using smart meter data. We designed a search space based on a novel channel attention fully convolutional network architecture. Furthermore, we developed a search algorithm based on Bayesian optimization to effectively explore the space and identify high-performing architectures. Results: The performance of the proposed method was evaluated and compared with a set of machine learning and deep learning baseline methods using a smart meter dataset widely used in this research area. Our results show that the deep neural network architectures designed automatically by our proposed method significantly outperform all baseline methods in addressing the socio-demographic questions investigated in our study. Full article

Convolutional neural networks (CNNs) are increasingly recognized as an important and potent artificial intelligence approach, widely employed in many computer vision applications, such as facial recognition. Their importance resides in their capacity to acquire hierarchical features, which is essential for recognizing complex patterns. Nevertheless, the intricate architectural design of CNNs leads to significant computing requirements. To tackle these issues, it is essential to construct a system based on field-programmable gate arrays (FPGAs) to speed up CNNs. FPGAs provide fast development capabilities, energy efficiency, decreased latency, and advanced reconfigurability. A facial recognition solution by leveraging deep learning and subsequently deploying it on an FPGA platform is suggested. The system detects whether a person has the necessary authorization to enter/access a place. The FPGA is responsible for processing this system with utmost security and without any internet connectivity. Various facial recognition networks are accomplished, including AlexNet, ResNet, and VGG-16 networks. The findings of the proposed method prove that the GoogLeNet network is the best fit due to its lower computational resource requirements, speed, and accuracy. The system was deployed on three hardware kits to appraise the performance of different programming approaches in terms of accuracy, latency, cost, and power consumption. The software programming on the Raspberry Pi-3B kit had a recognition accuracy of around 70–75% and relied on a stable internet connection for processing. This dependency on internet connectivity increases bandwidth consumption and fails to meet the required security criteria, contrary to ZYBO-Z7 board hardware programming. Nevertheless, the hardware/software co-design on the PYNQ-Z2 board achieved an accuracy rate of 85% to 87%. It operates independently of an internet connection, making it a standalone system and saving costs. Full article

In alignment with Saudi Vision 2030’s strategic objectives to diversify and enhance the tourism sector, this study explores the integration of Artificial Intelligence (AI) in the Al-Baha district, a prime tourist destination in Saudi Arabia. Our research introduces a hybrid AI-based framework that leverages sentiment analysis to assess and enhance tourist satisfaction, capitalizing on data extracted from social media platforms such as YouTube. This framework seeks to improve the quality of tourism experiences and augment the business value within the region. By analyzing sentiments expressed in user-generated content, the proposed AI system provides real-time insights into tourist preferences and experiences, enabling targeted interventions and improvements. The conducted experiments demonstrated the framework’s efficacy in identifying positive, neutral and negative sentiments, with the Multinomial Naive Bayes classifier showing superior performance in terms of precision and recall. These results indicate significant potential for AI to transform tourism practices in Al-Baha, offering enhanced experiences to visitors and driving the economic sustainability of the sector in line with the national vision. This study underscores the transformative potential of AI in refining operational strategies and aligning them with evolving tourist expectations, thereby supporting the broader goals of Saudi Vision 2030 for the tourism industry. Full article

WiFi-based human sensing has emerged as a transformative technology for advancing sustainable living environments and promoting well-being by enabling non-intrusive and device-free monitoring of human behaviors. This offers significant potential in applications such as smart homes and sustainable urban spaces and healthcare systems that enhance well-being and patient monitoring. However, current research predominantly addresses single-user scenarios, limiting its applicability in multi-user environments. In this work, we introduce “MultiSenseX”, a cutting-edge system leveraging a multi-label, multi-view Transformer-based architecture to achieve simultaneous localization and activity recognition in multi-occupant settings. By employing advanced preprocessing techniques and utilizing the Transformer’s self-attention mechanism, MultiSenseX effectively learns complex patterns of human activity and location from Channel State Information (CSI) data. This approach transcends traditional sequential methods, enabling accurate and real-time analysis in dynamic, multi-user contexts. Our empirical evaluation demonstrates MultiSenseX’s superior performance in both localization and activity recognition tasks, achieving remarkable accuracy and scalability. By enhancing multi-user sensing technologies, MultiSenseX supports the development of intelligent, efficient, and sustainable communities, contributing to SDG 11 (Sustainable Cities and Communities) and SDG 3 (Good Health and Well-being) through safer, smarter, and more inclusive urban living solutions. Full article

This study introduces the Eye-tracking Translation Software (ETS), a system that leverages eye-tracking data and real-time translation to enhance reading flow for non-native language users in complex, technical texts. By measuring the fixation duration, we can detect moments of cognitive load, ETS selectively provides translations, maintaining reading flow and engagement without undermining language learning. The key technological components include a desktop eye-tracker integrated with a custom Python-based application. Through a user-centered design, ETS dynamically adapts to individual reading needs, reducing cognitive strain by offering word-level translations when needed. A study involving 53 participants assessed ETS’s impact on reading speed, fixation duration, and user experience, with findings indicating improved comprehension and reading efficiency. Results demonstrated that gaze-based adaptations significantly improved their reading experience and reduced cognitive load. Participants positively rated ETS’s usability and were noted through preferences for customization, such as pop-up placement and sentence-level translations. Future work will integrate AI-driven adaptations, allowing the system to adjust based on user proficiency and reading behavior. The study contributes to the growing evidence of eye-tracking’s potential in educational and professional applications, offering a flexible, personalized approach to reading assistance that balances language exposure with real-time support. Full article

Background: COVID-19 genetic sequence research is crucial despite immunizations and pandemic control. COVID-19-causing SARS-CoV-2 must be understood genomically for several reasons. New viral strains may resist vaccines. Categorizing genetic sequences helps researchers track changes and assess immunization efficacy. Classifying COVID-19 genome sequences with other viruses helps to understand its evolution and interactions with other illnesses. Methods: The proposed study introduces a deep learning-based COVID-19 genomic sequence categorization approach. Attention-based hybrid deep learning (DL) models categorize 1423 COVID-19 and 11,388 other viral genome sequences. An unknown dataset is also used to assess the models. The five models’ accuracy, f1-score, area under the curve (AUC), precision, Matthews correlation coefficient (MCC), and recall are evaluated. Results: The results indicate that the Convolutional neural network (CNN) with Bidirectional long short-term memory (BLSTM) with attention layer (CNN-BLSTM-Att) achieved an accuracy of 99.99%, which outperformed the other models. For external validation, the model shows an accuracy of 99.88%. It reveals that DL-based approaches with an attention layer can accurately classify COVID-19 genomic sequences with a high degree of accuracy. This method might assist in identifying and classifying COVID-19 virus strains in clinical situations. Immunizations have lowered COVID-19 danger, but categorizing its genetic sequences is crucial to global health activities to plan for recurrence or future viral threats. Full article

(1) Background: Language represents a crucial ability of humans, enabling communication and collaboration. ChatGPT is an AI chatbot utilizing the GPT (Generative Pretrained Transformer) language model architecture, enabling the generation of human-like text. The aim of the research was to assess the effectiveness of ChatGPT-3.5 and the latest version, ChatGPT-4, in responding to questions posed within the scope of a periodontology specialization exam. (2) Methods: Two certification examinations in periodontology, available in both English and Polish, comprising 120 multiple-choice questions, each in a single-best-answer format. The questions were additionally assigned to five types in accordance with the subject covered. These exams were utilized to evaluate the performance of ChatGPT-3.5 and ChatGPT-4. Logistic regression models were used to estimate the chances of correct answers regarding the type of question, exam session, AI model, and difficulty index. (3) Results: The percentages of correct answers obtained by ChatGPT-3.5 and ChatGPT-4 in the Spring 2023 session in Polish and English were 40.3% vs. 55.5% and 45.4% vs. 68.9%, respectively. The periodontology specialty examination test accuracy of ChatGPT-4 was significantly better than that of ChatGPT-3.5 for both sessions (p < 0.05). For the ChatGPT-4 spring session, it was significantly more effective in the English language (p = 0.0325) due to the lack of statistically significant differences for ChatGPT-3.5. In the case of ChatGPT-3.5 and ChatGPT-4, incorrect responses showed notably lower difficulty index values during the Spring 2023 session in English and Polish (p < 0.05). (4) Conclusions: ChatGPT-4 exceeded the 60% threshold and passed the examination in the Spring 2023 session in the English version. In general, ChatGPT-4 performed better than ChatGPT-3.5, achieving significantly better results in the Spring 2023 test in the Polish and English versions. Full article

This study investigates the use of machine learning (ML) models combined with explainable artificial intelligence (XAI) techniques to identify the most influential genes in the classification of five recurrent cancer types in women: breast cancer (BRCA), lung adenocarcinoma (LUAD), thyroid cancer (THCA), ovarian cancer (OV), and colon adenocarcinoma (COAD). Gene expression data from RNA-seq, extracted from The Cancer Genome Atlas (TCGA), were used to train ML models, including decision trees (DTs), random forest (RF), and XGBoost (XGB), which achieved accuracies of 98.69%, 99.82%, and 99.37%, respectively. However, the challenges in this analysis included the high dimensionality of the dataset and the lack of transparency in the ML models. To mitigate these challenges, the SHAP (Shapley Additive Explanations) method was applied to generate a list of features, aiming to understand which characteristics influenced the models’ decision-making processes and, consequently, the prediction results for the five tumor types. The SHAP analysis identified 119, 80, and 10 genes for the RF, XGB, and DT models, respectively, totaling 209 genes, resulting in 172 unique genes. The new list, representing 0.8% of the original input features, is coherent and fully explainable, increasing confidence in the applied models. Additionally, the results suggest that the SHAP method can be effectively used as a feature selector in gene expression data. This approach not only enhances model transparency but also maintains high classification performance, highlighting its potential in identifying biologically relevant features that may serve as biomarkers for cancer diagnostics and treatment planning. Full article

Background: Integrating nonlinear behavior into the architecture of artificial neural networks is regarded as essential requirement to constitute their effectual learning capacity for solving complex tasks. This claim seems to be true for moderate-sized networks, i.e., with a lower double-digit number of layers. However, going deeper with neural networks regularly turns into destructive tendencies of gradual performance degeneration during training. To circumvent this degradation problem, the prominent neural architectures Residual Network and Highway Network establish skip connections with additive identity mappings between layers. Methods: In this work, we unify the mechanics of both architectures into Capsule Networks (CapsNet)s by showing their inherent ability to learn skip connections. As a necessary precondition, we introduce the concept of Adaptive Nonlinearity Gates (ANG)s which dynamically steer and limit the usage of nonlinear processing. We propose practical methods for the realization of ANGs including biased batch normalization, the Doubly-Parametric ReLU (D-PReLU) activation function, and Gated Routing (GR) dedicated to extremely deep CapsNets. Results: Our comprehensive empirical study using MNIST substantiates the effectiveness of our developed methods and delivers valuable insights for the training of very deep nets of any kind. The final experiments on Fashion-MNIST and SVHN demonstrate the potential of pure capsule-driven networks with GR. Full article