Expert System Research Articles

Can a rule-based expert system diagnose nasal obstruction from nasoendoscopy videos?

Nasal obstruction has multiple causes requiring specialist endoscopy for diagnosis. A rule-based expert system (RB-ES), which applies five 'if-then' rules based on nasal features, may help replicate ENT decision-making in settings with limited access. Objectives & Hypotheses: This study evaluated RB-ES in diagnosing allergic rhinitis (AR), chronic rhinosinusitis with (CRSwNP) and without (CRSsNP) nasal polyps, and deviated nasal septum (DNS). Primary outcomes were sensitivity and specificity; secondary outcome was agreement with ENT specialists. Prospective cohort study Methods: Seventy-one participants (65 patients, 6 controls) underwent pre- and post-decongestion endoscopy. Four ENT specialists provided diagnoses. RB-ES performance was compared against confirmed clinical diagnoses. RB-ES showed no detectable significant sensitivity differences from ENT specialists (all p > 0.05). Sensitivity was highest for CRSwNP; specificity remained high overall. RB-ES matched specialist performance in CRSwNP diagnosis. Dataset expansion and AI integration are recommended for further validation.

A transformer based generative chemical language AI model for structural elucidation of organic compounds

For over half a century, computer-aided structural elucidation systems (CASE) for organic compounds have relied on complex expert systems with explicitly programmed algorithms. These systems are often computationally inefficient for complex compounds due to the vast chemical structural space that must be explored and filtered. In this study, we present a proof-of-concept transformer based generative chemical language artificial intelligence (AI) model, an innovative end-to-end architecture designed to replace the logic and workflow of the classic CASE framework for ultra-fast and accurate spectroscopic-based structural elucidation. Our model employs an encoder-decoder architecture and self-attention mechanisms, similar to those in large language models, to directly generate the most probable chemical structures that match the input spectroscopic data. Trained on ~ 102 k IR, UV, and 1H NMR spectra, it performs structural elucidation of molecules with up to 29 atoms in just a few seconds on a modern CPU, achieving a top-15 accuracy of 83%. This approach demonstrates the potential of transformer based generative AI to accelerate traditional scientific problem-solving processes. The model's ability to iterate quickly based on new data highlights its potential for rapid advancements in structural elucidation.

Graphical Empirical Mode Decomposition–Convolutional Neural Network-Based Expert System for Early Corrosion Detection in Truss-Type Bridges

Corrosion is a critical issue in civil structures, significantly affecting their durability and functionality. Detecting corrosion at an early stage is essential to prevent structural failures and ensure safety. This study proposes an expert system based on a novel methodology for corrosion detection using vibration signal analysis. The approach employs graphical empirical mode decomposition (GEMD) to decompose vibration signals into their intrinsic mode functions, extracting relevant structural features. These features are then transformed into grayscale images and classified using a Convolutional Neural Network (CNN) to automatically differentiate between a healthy structure and one affected by corrosion. To enhance the computational efficiency of the method without compromising accuracy, different CNN architectures and image sizes are tested to propose a low-complexity model. The proposed approach is validated using a 3D nine-bay truss-type bridge model encountered in the Vibrations Laboratory at the Autonomous University of Querétaro, Mexico. The evaluation considers three different corrosion levels: (1) incipient, (2) moderate, and (3) severe, along with a healthy condition. The combination of GEMD and CNN provides a highly accurate corrosion detection framework that achieves 100% classification accuracy while remaining effective regardless of the damage location and severity, making it a reliable tool for early-stage corrosion assessment that enables timely maintenance and enhances structural health monitoring to improve the long life and safety of civil structures.

CREATION OF A MATHEMATICAL MODEL OF A SYSTEM FOR EARLY DIAGNOSIS OF LUNG CANCER

Oncological diseases are one of the deadliest diseases in the modern world. Early diagnosis of these diseases helps to increase the life expectancy of patients. The development of measures and programs for the early diagnosis of oncological diseases is one of the urgent problems today. The use of intelligent systems and artificial intelligence methods in the diagnosis of cancer is an important aspect in this matter. In such cases, expert systems or decision support systems are mainly used. This paper examines the early diagnosis of lung cancer using a protocol survey and taking into account regional factors. The research is conducted for residents of the former Semipalatinsk nuclear test site. As the tests carried out have an impact on the health of the citizens of this region to this day. The region is one of the five regions with the highest incidence of oncological diseases and mortality rates. An analysis of the existing expert systems was carried out. Neural networks for decision support systems were used as the basis of the model. Each parameter of the model is assigned a weight, relative to which the significance is calculated and preliminary instructions are given on the further actions of the interviewed patient. As a result, the factors that most strongly influence the incidence of lung cancer were identified.

EXPLORING THE LANDSCAPE OF EXPERT SYSTEMS: A REVIEW

Expert systems are computer programs designed to mimic the decision-making of human experts. This paper explores the fundamental components of ES, including knowledge acquisition, representation, and reasoning. Various techniques for acquiring knowledge, such as interviews, observation, and document analysis, are discussed, along with prominent knowledge representation schemes like production rules, semantic networks, frames, and ontologies. The reasoning process, including inference methods and explanation facilities, is also examined. The paper further analyses the challenges and limitations of ES, such as the difficulty in capturing common sense reasoning and the complexity of knowledge base maintenance. Finally, it explores future research directions, including the integration of emerging technologies like big data and cloud computing, the development of more transparent and explainable ES, and addressing ethical considerations surrounding bias and accountability. This comprehensive overview provides a foundational understanding of expert systems, their capabilities, limitations, and potential future advancements.

CONCEPTUAL FOUNDATIONS OF THE EXPERT SYSTEM OF BIOCLIMATIC MODELING

The article is devoted to the theoretical substantiation of the structure of the Expert System for Bioclimatic Modeling and the disclosure of its functioning mechanisms. The main goal is to develop the conceptual framework of an expert system as an innovative tool to support analytical and creative decision-making in architectural design. The system is designed to integrate a variety of multi-parameter data, including climatic conditions, characteristics of architectural objects, etc. to generate optimal bioclimatic recommendations. In the course of the study, a set of scientific methods was used to develop the structure of the expert system. System analysis was used to identify the key components of the system and their interrelationships. Cluster analysis was used to develop the structure of the knowledge base for effective organization and systematization of information. Information system modeling techniques were used to develop a model of interaction between the expert system modules, including the database, rule base, and solver. Expert evaluation methods were used to fill the knowledge base and evaluate the parameters of design decisions. A multi-criteria approach was used to process fuzzy data and make decisions in the face of the complexity and multiplicity of bioclimatic modeling criteria. The results of the study present the conceptual structure of the expert system, where a hybrid knowledge base combining declarative and procedural knowledge, a modular database structure, the use of the heuristic method "IF-THET" in the rule base, and a hierarchical system based on a decision tree are proposed to effectively process complex information and generate optimal bioclimatic recommendations. Further research is planned to focus on the practical implementation and validation of the developed system, expanding its functionality by integrating with other design tools, and using machine learning methods to automatically update the knowledge base.

Mapping the AI Landscape in Food Science and Engineering: A Bibliometric Analysis Enhanced with Interactive Digital Tools and Company Case Studies

Abstract The proliferation of research on Artificial Intelligence (AI) in food science and engineering has made it increasingly difficult to synthesise relevant insights effectively. Although AI adoption in the food industry has grown, it lags behind sectors like finance and healthcare due to the complexity of food systems, including high process variability, risk aversion towards novel technologies, and constrained investment appetite. Historically, computational techniques and AI-adjacent technologies like expert systems and empirical modelling have supported food research and development for decades. More recently, AI applications have broadened to include process control, food safety, ingredient and product quality, sensory evaluation, traceability, and supply chain management. In response to the rapid increase in AI-related food science publications – particularly since 2019 – this review introduces tools for dynamically synthesising and exploring this evolving knowledge base. We present an interactive dashboard that integrates a curated dataset of food AI review articles with advanced bibliometric analyses, enabling user-driven exploration of research trends and thematic relationships. Additionally, we demonstrate the use of customised large language model (LLM) tools for targeted literature interrogation, enhancing accessibility for researchers and industry stakeholders. Complementing this academic synthesis, we profile selected industry case studies where AI plays a central role in ingredient discovery, product development, intelligent sorting, and sensory analytics. By combining interactive research tools with real-world case studies, this review offers a comprehensive snapshot of Food AI and begins to bridge the gap between academic research and industry implementation, providing a valuable resource for those seeking both domain-specific knowledge and actionable insights.

Harnessing AI for Improved Diagnosis and Management of Pediatric Sepsis: Current Advances, Challenges, and Future Directions.

Artificial intelligence (AI) has been applied to early recognition and management of rapidly progressive, community-acquired pediatric sepsis, a leading cause of childhood mortality. The broad adoption of electronic health records combined with rapid advances in digital technologies have enabled the federated training of both knowledge-driven AI, known as expert systems, trained by teams of collaborating clinicians, and data-driven AI, known as machine learning (ML), to derive predictive, clustering algorithms trained on "big data." An important subset of ML is "deep learning," which includes tools that understand, interpret, and manipulate human imagery and language, such as natural language processing and its subset large language models. We are in an era of rapid deployment of AI/ML-powered tools ranging from real-time electronic health records-embedded decision support tools to continuous wearable vital sign monitors and mobile/conversational virtual assistants/triage apps. These applications have the potential of transforming the timeliness of life-saving sepsis care delivery. This review explores the current and potential AI/ML applications in sepsis care, including tools for screening/early detection, risk stratification/outcome prediction, personalized treatment, and continuous patient monitoring. We highlight successful implementations and ongoing clinical trials, emphasizing the impact on patient outcomes. Finally, we address practical considerations for the future, such as bias mitigation and integration into clinical workflows.

Expert decision support system for biologically inspired product design integrating emotional preferences and image generation

Expert decision support system for biologically inspired product design integrating emotional preferences and image generation

The missions of a regional system of expertise and support for the organization of the adult obesity network

The missions of a regional system of expertise and support for the organization of the adult obesity network

Expert system for extracting keywords in educational texts and textbooks based on transformers models

Expert system for extracting keywords in educational texts and textbooks based on transformers models

A mini-review of the development and use of expert systems in mining

Purpose. The mining industry is one of the sectors that have benefited from expert systems over the years. This review aims to analyze developments in using expert systems in mining. Methods. The approach used involved searching, screening, and selecting relevant published studies following the Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) methodology. A total of 32 (n = 32) articles reporting on specific expert systems developed for the mining industry were considered for further analysis while 21 (n = 21) were excluded from the study. The analysis looked at the nature of the reported expert systems, their mining application areas, and the tools used to develop them. Findings. The results reveal that there is generally an increase in the development and use of expert systems in the mining sector. The abundant availability of expert system shells and the adoption of recent digital technologies such as cloud computing and the Internet of Things (IoT) present a potential for further development of expert systems in the mining industry. Originality. This is the first review of the trends in the development and use of expert systems in mining. Practical implications. This work’s findings give insights into the trends and opportunities for the development and application of expert systems in mining. The growing use of expert systems in making sound decisions in mining has the potential to make future mining operations safe, profitable, and sustainable.

The Effect of Artificial Intelligence on Strategic Decision-Making in Multinational Corporations

This study aimed to identify the impact of artificial intelligence on the strategic decision-making process in multinational companies. The study used the descriptive analysis method and collected data using a random sample. The study sample consisted of 456 individuals. The research reached a number of results, one of which is that a large number of multinational companies in Saudi Arabia use computers equipped with artificial intelligence, and it has been proven that these institutions are capable of making strategic decisions. In addition, a high level of application of intelligent agents, artificial neural networks, and expert systems has been proven. In addition, intelligent agents, expert systems, artificial neural networks, and artificial intelligence have all been shown to have a significant impact on the process by which multinational companies make their strategic decisions. The study concluded that enabling multinational companies in Saudi Arabia to fully benefit from artificial intelligence requires creating the appropriate environment for its application and providing the necessary technological infrastructure. The study also emphasized the importance of employing individuals with extensive scientific and practical knowledge of artificial intelligence, and sought to provide Saudi employees with training programs and courses to maximize their potential in dealing with artificial intelligence and related technologies.

The Impact of Artificial Intelligence Applications on the Quality of Accounting Information

This research aims to investigate the impact of Artificial Intelligence Applications (AIA), with its dimensions—expert systems, neural networks, genetic algorithms, and intelligent agents—on the quality of accounting information. The research methodology employed a mixed-method approach, combining both quantitative and qualitative techniques. A survey was conducted among accounting and auditing professionals to collect data on the adoption of AI applications and their perceived impact on the quality of accounting information. Statistical analysis methods, including correlation and regression analysis, were applied to examine the relationships between the adoption of AI applications and the targeted outcomes. The study yielded several findings, most notably that the use of genetic algorithms represents one of the most significant and influential applications of artificial intelligence in enhancing the quality of accounting information. The results indicated no statistically significant effect of expert systems, neural networks, or intelligent agents on the quality of accounting information. However, the genetic algorithms dimension had a positive and significant impact. This may be attributed to the professional and technological environment in Yemen, which has not yet reached the maturity required to effectively benefit from more complex AI technologies such as expert systems, neural networks, and intelligent agents. In contrast, the environment has benefited, at least partially, from more practical and effective analytical tools such as genetic algorithms, as reflected in the study's findings. The study concludes with several recommendations, most importantly the need to adopt training programs and workshops for chartered accountants to enhance their understanding and application of artificial intelligence technologies

Integrating Gamification in Expert Systems: A Novel Approach for Stress Disorder Diagnosis in Digital Mental Health

The increasing prevalence of stress disorders highlights the need for innovative, accessible, and engaging diagnostic tools in mental health services. This study presents the design and implementation of a gamified expert system for diagnosing stress disorders, integrating gamification elements to enhance user engagement and reduce stigma. The system employs the forward chaining method to deliver high-accuracy, rule-based diagnoses while incorporating features such as points, rewards, and leaderboards to motivate user interaction.The system's development followed a user-centered design approach to ensure an intuitive interface aligned with user needs. Evaluation results demonstrated a diagnostic accuracy rate of 92%, validated by mental health professionals, alongside significant improvements in user engagement metrics, including session frequency and duration. Qualitative feedback indicated that gamification effectively reduced stigma and increased motivation for mental health assessments.These findings suggest that gamified expert systems can bridge gaps in accessibility and engagement in mental health services. This research contributes to the advancement of digital health technologies by providing practical insights into integrating gamification into expert systems to foster proactive mental health management.

РОЗРОБЛЕННЯ СИСТЕМИ НЕЧІТКОГО ЛОГІЧНОГО ВИСНОВКУ ПРИЙНЯТТЯ УПРАВЛІНСЬКИХ РІШЕНЬ У ВИПРОБУВАЛЬНИХ ЛАБОРАТОРІЯХ

The purpose of the article is to develop a fuzzy logic inference system for making management decisions in testing laboratories. In accordance with the requirements of DSTU EN ISO/IEC 17025:2019, the risk and capability management system shall be an element of the laboratory management system. To optimize the risk management process, an expert system based on fuzzy modeling of the risk significance assessment output was used. In the process of analyzing risk factors, indicators were identified that can be defined as risk indicators for testing laboratories: calibration interval of the applied measuring equipment; calibration result of the applied measuring equipment; number of errors in product test protocols; result of participation in comparative tests; time allocated for product testing. The model contains rule base and allows for linguistic analysis of risks for the activities of a testing laboratory. The developed rule base contains 243 rules. The main rule: when one of the risks is high, then the result also has a high risk. The implementation of the rule base modeling process is carried out using the specialized Fuzzy Logic Toolbox package of the MATLAB software. The fuzzy inference is implemented based on the Mamdani algorithm.. The resulting surface of the dependence of the output linguistic variable on two input risk factors: the calibration interval of the applied measuring equipment and the calibration result of this measuring equipment. The surface clearly demonstrates the validity and consistency of the developed system of production rules. The defuzzification procedure was carried out using the "center of gravity" method, the results for several combinations of input data are presented in the table. The results of this modeling process allow determining the need for preventive measures or corrective actions. The presence of the specified expert system for determining the significance of risks in testing laboratories will ensure constant monitoring of the state of the risk management system. A further perspective for the development of the proposed model is the creation of an adaptive fuzzy production model, which will allow for their reassessment when new risks arise.

Implementasi Algoritma Genetika untuk Pengaturan Gizi Remaja

Adolescence is an important moment in human life marked by growth, emotional, and psychosocial. During adolescence, a healthy diet becomes very crucial to support adolescent development and prevent future health problems. Therefore, this study aims to implement the Genetic Algorithm for Teenager Nutrition Management in an expert system. This system is designed to provide recommendations for food menus that are in accordance with the nutritional needs of a teenager, which include protein, carbohydrates, energy/calories, and fat. Genetic algorithms are used so that the food recommendation process can be in accordance with the daily nutritional needs of teenagers. The variables in this study are age, gender, weight, height, and physical activity. Meanwhile, the results of this system are in the form of total energy or calorie needs, protein, fat, and carbohydrates included in the daily food menu according to the nutritional needs of teenagers. This system can provide daily food menu recommendations in the form of a menu list using a genetic algorithm. The best fitness value is 0.0098 at a generation size of 100 and 670 dataset of food

Advances in Transformer-Based Semantic Search: Techniques, Benchmarks, and Future Directions

Semantic search has developed quickly as the need for accurate information retrieval has increased in a variety of fields, from expert knowledge systems to web search engines. Conventional search methods that rely solely on keywords frequently fail to understand user intent and contextual hints. This survey focuses on recent advances in Transformer-based models, such as BERT, RoBERTa, T5, and GPT, which leverage self-attention mechanisms and contextual embeddings to deliver heightened precision and recall across diverse domains. Key architectural elements underlying these models are discussed, including dual-encoder and cross-encoder frameworks, and how Dense Passage Retrieval extends their capabilities to large-scale applications is examined. Practical considerations, such as domain adaptation and fine-tuning strategies, are reviewed to highlight their impact on real-world deployment. Benchmark evaluations (e.g., MS MARCO, TREC, and BEIR) are also presented to illustrate performance gains over traditional Information Retrieval methods and explore ongoing challenges involving interpretability, bias, and resource-intensive training. Lastly, emerging trends—multimodal semantic search, personalized retrieval, and continual learning—that promise to shape the future of AI-driven information retrieval are identified for more efficient and interpretable semantic search.

Application of artificial intelligence in forecasting corporate financial risks

In the context of increasing economic uncertainty, climate variability, and rising credit exposure in Kazakhstan’s key industries, the ability to accurately forecast corporate financial risks has become critically important. This article examines the practical application of artificial intelligence (AI) in improving financial risk prediction, with a focus on Kazakhstan’s agricultural and leasing sectors. The study evaluates how user-friendly AI tools—such as rule-based expert systems, satellite imagery analytics, automated reporting modules, and AI-powered inventory platforms—can support early warning systems and enhance operational decision-making. Scientific relevance is ensured through the analysis of real-life cases, including Farmonaut’s crop monitoring system and KazAgroFinance’s AI-driven asset verification tool. These technologies enable more accurate forecasts of yields and drought risks, improve resource planning, reduce manual errors, and strengthen loan oversight. The methodology is based on comparative analysis of financial and operational indicators over a 10-year period (2015–2024), with a focus on pre- and post-AI implementation outcomes. Unlike complex predictive models, the study emphasizes transparent, easily interpretable AI applications that are already in use in Kazakhstan. The findings confirm that such tools enhance efficiency, mitigate financial risks, and increase resilience in corporate management. This research offers practical recommendations for scaling up AI-based forecasting tools in Kazakhstan’s business ecosystem.

Fuzzy Expert System for Decission Support to Diagnosis Leukemia

Leukemia is a cancer of the blood and bone marrow. In leukemia, the bone marrow produces too many abnormal white blood cells. These abnormal cells cannot fight infections well and can displace healthy blood cells, which can cause anemia and bleeding. In this study, a fuzzy method will be implemented to diagnose leukemia and the results will later be compared with expert diagnoses. Fuzzy logic was chosen because it allows for degrees of truth between 0 (completely false) and 1 (completely true) and it is suitable for situations where human expertise relies on experience and judgment rather than fixed rules. Fuzzy systems can analyze large amounts of data quickly, thereby accelerating the diagnosis and decision-making process, especially when used in medical decision support systems. This study produced a leukemia diagnosis accuracy of 88.83% when compared with the results of expert diagnoses using the same symptom and sample data.