10th International Congress on Information and Communication Technology in concurrent with ICT Excellence Awards (ICICT 2025) will be held at London, United Kingdom | February 18 - 21 2025.
Sign up or log in to bookmark your favorites and sync them to your phone or calendar.
Authors - Varsha Naik, Rajeswari K, Kshitij Jadhav, Aniket Rahalkar Abstract - This study examines cross-lingual natural language processing (NLP) techniques to address the challenges of developing conversational AI systems for low-resource languages. These languages often lack extensive linguistic re- sources such as large-scale corpora, annotated datasets, and language-specific tools, making it difficult to capture the linguistic distinctions and contextual meaning essential for high-quality dialogue systems. This language gap restricts accessibility and inclusivity, preventing speakers of these underrepresented languages from fully benefiting from advancements in technology. The study compares various factors that affect model performance, including transformer model architecture, cross-lingual embeddings, fine-tuning strategies, and transfer learning approaches. Despite these challenges, the research shows that cross-lingual models offer promising solutions, especially when utilizing techniques like transfer learning and multilingual pre-training. By transferring knowledge from high-resource languages, these models can compensate for the scarcity of data in low-resource languages, enabling the development of more accurate, culturally sensitive, and inclusive AI systems. The findings highlight the importance of bridging linguistic divides to foster greater language diversity, accessibility, and technological inclusivity, ultimately supporting cultural preservation and revitalization.
Authors - Elrasheed Ismail Mohommoud Zayid, Ahmad Mohammad Aldaleel, Omar Abdullah Omar Alshehri Abstract - Machine learning classifiers are the first candidate methodology that could be used to assess the digital innovation across a set of teachers. This study aims to collect, build, represent, and discuss a reliable digital innovation skills (DIS) dataset by recruiting teachers chosen from the teachers who work in Bisha Province, Saudi Arabia. The study processed a rich data sample and made it accessible and shareable for the researchers' open use. DIS assessment addressed the problems and helped design a suitable innovation training module for the local community teachers. The total dataset comprises 400 conveniently collected data points, and each data point represents a complete record of teachers among the DSTs of Bisha Province. The research fields are prepared and set as fifty questionnaire questions, which distributed across the DSTs community in the area using social networks. Each question represents a single input or output feature for the classification model. Before running the ML models, the input variables are encoded serially from F0 to F49, and based on an explanatory test performed using LazyPredictools, only the positively contributing features are used. The extensive dataset, which is kept in the Mendeley Data repository, has a great deal of possibilities for reuse in sensitivity analysis, policymaking, and additional study. The decision tree, extra tree, and extreme gradient boosting (XGB) classifiers are examples of the recruited algorithms for evaluating DISs. The authors believe that this a wealthy kind of innovative respiratory dataset with its classification features will become a valuable mining source for interested researchers.
Authors - Malgorzata Pankowska Abstract - Business information system (BIS) consultants are working on solving problems of client companies, providing them with high-quality services, helping them quickly respond to changes in their ecosystems, and to the changes initiated by new technologies. Client is usually the most important actor in the consulting process. Therefore, the consultants are to be well educated to ensure the best satisfying solutions. This study focuses on business information system analysts’ competences development to enable them participation in the consulting projects. In this study, the thematic review of literature was applied, the author’s framework of consultants’ competencies for business information system strategic analysis has been provided, and finally, the author formulate a recommendation on business analysis course for students of computer science at university. The findings indicate that both the students’ motivation, knowledge, experience, as well as a strong theoretical background and a methodological support from cooperative business units influence innovativeness and creativity of BIS consultants.
Authors - Martin Mayembe, Jackson Phiri Abstract - Religious organizations, particularly church organisations, play a significant role in the lives of many people globally. These organisations require efficient management of various operations such as management of members, finances, events, and communications to fulfil their mission effectively. Existing church management systems are often built using traditional monolithic architectures, which come with inherent challenges. These challenges include platform dependence, limited scalability, and high upfront investment, making it difficult for many church organizations to develop, maintain, and scale their systems effectively and efficiently. This method of development is often referred to as the Spaghetti model. This study explores the application of the Micro-services Architecture in Church Management Systems, with the use of a service bus to enable communication between the services, to achieve modularity and scalability. To demonstrate the effectiveness of this design, a prototype is developed, focusing on two key modules: the Church Member Management System and the Financial Management System These modules work in tandem to manage member and associated member contribution data to provide access to up-to-date vital information.
Authors - Williams A. Ayara, Adenike O. Boyo, Mustapha O. Adewusi, Razaq O. Kesinro, Mojisola R. Usikalu, Kehinde D. Oyeyemi Abstract - The search for enhancing green electricity generation and the constant increase in the price of crude oil and its products propelled the choice of this research. Hence, a photovoltaic fuel-less power generating system using locally available materials. The input and output characteristics are analyzed to determine the efficiency, and the power generated by the photovoltaic-powered fuel-less generator is used to power an external load. The photovoltaic used is oriented to face in a direction with optimum tilt for maximum yield (to face southward) of solar power. This orientation and angle of tilt were determined using the Garmin Oregon450 GPS in conjunction with a Seaward Solar Survey 200R meter. Thus, the photovoltaic fuel-less generator was successfully developed. The driving component of this power-generating system is the 1 HP Direct Current (DC) motor, powered by two (2) 250 W mono-crystalline solar panels via a 12 V battery connected to a 30 A charge controller to maintain the charge level of the battery which helps to spin the 650 W Alternating Current (AC) alternator to deliver electricity. The device efficiently delivered power by lighting three (3) incandescent bulbs and a standing fan with total power between 100 – 220 W, and an efficiency of 70 -75%. This generator is eco-friendly since it does not emit any contaminants to the environment.
Authors - Sarah Anis, Mohamed Mabrouk, Mostafa Aref Abstract - This research paper investigates the application of sentiment analysis in the fintech sector, focusing on stock market prediction through a transformer-based model, specifically FinBERT. By comparing its performance against established models like CNN, LSTM, and BERT across different datasets, the study demonstrates that FinBERT achieves superior accuracy in classifying sentiments from financial reviews. The findings emphasize the significance of specialized models tailored to specific domains for improving sentiment analysis within the financial sector, providing useful information for those involved in the fintech field.
Authors - Carlo Giuseppe Pirrone Abstract - This study explores urban accessibility for the elderly, focusing on the importance of integrating objective and subjective measures for a comprehensive assessment. Objective measures, such as cumulative opportunity measures (CUM) or measurable travel times, quantify spatial data while often neglecting personal experiences and user perceptions. Subjective measures, obtained through surveys, become crucial in defining the ease of access to services, including satisfaction levels derived from the journey, barriers due to individual factors (age, health, disability), as well as comfort and safety. A combined methodology would promote a new interpretation of urban accessibility. A case study conducted in Rende, Italy, illustrates a practical application by mapping healthcare services and public transport to assess pedestrian accessibility. A pilot survey gathered elderly residents' perceptions of distance, travel time, and service satisfaction. Preliminary results indicate a reluctance to walk, overestimated perceived distances and a strong reliance on private vehicles, highlighting the need for infrastructures and services to provide a better connection of the elderly to healthcare services. Ongoing research will further refine the study by adapting objective measures to local perceptions to develop a specific accessibility indicator for the area.
Authors - Dilshan De Silva, Dulaj Dawlagala Abstract - The research demonstrates a reliable and efficient IoT solution aimed good at improving the safety of campers and hikers in outdoor settings. It consists of one primary device and a set of other peripherals, which employ LoRa communication, GPS positioning, and environmental parameters for location-based services. The primary device system consists of an ESP32 microcontroller fitted with a LoRa 433 MHz module and NEO-7MGPS powered using SYN-ACK protocol that allows the device to constantly communicate with the subordinate devices. The subordinate devices which are also based on ESP32 has LoRa modules and OLED screen for the purposes of receiving and providing information about the geofences and locational alerts. A significant strength of this system is that it is able to function in remote areas by bringing all the devices together into a mesh network so that data and devices can be synchronized without relying on the internet for connection purposes. Both primary and subordinate devices have the ability to connect to the internet wherever possible, update messages, synchronize, and transfer messages efficiently. Failure is further minimized for effective communication and precise positioning which are important in the management of outdoor safety hazards. The first prototype tests have shown the ability of the system to solve problems such as real-time interaction, data integration, and functioning in difficult environments. To npm and deploy the research further developed IoT systems for outdoor applications, effective outdoor deployment strategies were developed. As the next step, the system will be tested on a larger scale to assess its scalability, and user-centered interfaces will be redesigned to accommodate real-world scenarios.
Authors - Jude Osakwe, Sinte Mutelo, Nelson Osakwe Abstract - This study aimed to investigate the interrelationship between data governance and compliance with privacy regulations. A systematic literature review was conducted to synthesise the existing research on data governance's impact on compliance with privacy regulations. The study found that data governance has a positive association with compliance, with integrated data governance methods and processes supporting decision-making, and stakeholders' involvement guaranteeing transparent processes. The findings also suggested that the impact of data governance on privacy regulations compliance needs a certain maturity level and top management support. Key recommendations for organisations are outlined to enhance their governance frameworks, promote transparency, and align resources effectively in order to bolster compliance with privacy regulations. The study concludes by addressing identified research gaps and offering directions for future studies aimed at exploring the evolving landscape of data governance and privacy compliance.
Authors - Nabeela Kausar, Ramiza Ashraf, Saeed Nawaz Khan Abstract - Skin cutaneous melanoma is one of the most aggressive forms of skin cancer, with various prognostic factors that can significantly impact patients’ survival outcomes. Survival analysis helps in identifying key factors influencing patient outcomes and guides in clinical decision-making. In literature, statistical methods have been used for the survival analysis of skin cancer patients but these methods have limitations. To address the limitations of traditional statistical methods in survival analysis, researchers have developed a range of machine learning (ML) based survival analysis techniques. These ML techniques offer advanced capabilities for modeling complex relationships and improving prediction accuracy. But "black box" nature of ML models poses a challenge, especially in fields like healthcare where understanding the rationale behind predictions is crucial. It this work, Explainable AI (XAI) based survival analysis has been carried out using XGboost model and clinical features of skin cutaneous melanoma patients. XAI models explain their prediction by showing the important features involved in the prediction to demonstrate their reliability to be used by the clinicians. To validate the performance of XAI model, in this work, multivariate regression based Cox Proportional Hazard (CPH) model has been developed which shows the relationship of patients’ clinical features and survival time. The pro-posed XAI based model has C-index value of 84.3% and shows that age, pathology T stage, and pathology N stage are key factors influencing the survival of skin cutaneous melanoma patients. The CPH model further validates the strong association between these features and patient survival.
Authors - Mulima Chibuye, Jackson Phiri Abstract - Agricultural systems have been modeled and prediction of yields used in the space since the beginning of agriculture ,improvements in the crop science and better tools made the task much easier through the ages, from using the position of the sun to determining that certain weeds signify a good harvest to actually determining what factors precede observable phenomena, the space has be-come so advanced such that we are able to build better prediction models and the promise of quantum computation that can model much more complex systems and interactions among the individual parameters within the system promise to make us predict yields of crops with much better accuracy than has ever been deemed feasible. With the technology that we have available now, we can apply properties of physical systems on classical computers such as mimicking chaos theory to add randomness to our predictions as that is the way nature works. That randomness is due to how initial conditions might potentially fluctuate and we would normally call it random because we are missing certain parameters that if we collect, would greatly improve how we predict physical chaotic systems. The aim of this work is to explore how we can incorporate chaos in agricultural systems by making use of a hybrid approach to known systems like dense neural networks and more recent methods such as Echo State Networks.
Authors - Cynthia Sherin B, Poovammal E Abstract - Vehicle detection and classification has a key role in evolution of intelligent transportation system. Accuracy in the detection enhances the efficiency of intelligent traffic monitoring systems. The paper shows a comparative study on the performance of YOLOv8 and YOLOv10 detection models on the novel Vehicle Identification aNd Classification (VINC) dataset introduced. They detect multi-class vehicles such as cars, trucks, buses, bicycles and bikes. The achievements of each model are assessed using precision, recall, F1 score and confusion matrices. The experimental results demonstrates the supremacy of YOLOv10 in detecting very small and more complex vehicle structures in the traffic scenario than YOLOv8. Alternatively, YOLOv8 also exhibited equivalent detection accuracy in detecting large vehicles like buses and trucks, by capturing the minute variations in the processed features. The detection models achieve precision of 97.2% and 93.6% for YOLOv10 and YOLOv8 respectively. YOLOv10 achieves high recall rate and F1 score of 92.4% and 81.4% respectively. Thus, the detection performance of these vehicles expresses the robust characteristics of both the YOLO versions. This research paper delineates the merits and drawbacks of these two versions on real-time circumstances, thereby creating faster and precise detection models.
Authors - Mariam Basim Al-Najjar, Khalil H. Sayidmarie Abstract - This contribution investigates a proposed Frequency Selective Surfaces (FSS) that can be reconfigured using light-dependent resistor LDR. The unit cell of the FSS comprises a split square ring equipped with a single LDR placed at its gap. The FSS is built on the FR4 substrate of 40X40 mm dimensions, and ring size of 29 x29 mm to serve the WLAN application of 2.45 GHz frequency. When the LDR is adequately illuminated it exhibits a small resistance, and the ring behaves as a closed one, while in the dark condition, the resistance is high and the ring acts as a split ring. Therefore, the FSS works as a bandpass filter when illuminated, and as a bandstop filter without illumination. The LDR doesn’t need biasing wires that usually interfere with the structure of the FSS.
Authors - Moubaric Kabore, Abdoulaye Sere, Vini Yves Bernadin Loyara Abstract - This paper deals with Bayesian approach in Data Research in GIS database through artificial Intelligence (AI) modules, reading the best bayesian probability before returning the data requested, denoted AI4DB. The proposed method combines meshing techniques and the map-reduce algorithm with Bayesian approach to obtain a smart GIS database to reduce the execution time. According to the values of the Bayesian probability, the nearest sites of any position resulting of the user requests, are extracted speedily from the database using the map reduce framework. The execution time is less than the time for the case of the classical method, based only on a parallelism search without a probability. Only a map function with the best bayesian probability for the data in entry, executes entirely its instruction.
Authors - Denis Vasiliev, Rodney Stevens, Lennart Bornmalm Abstract - The idea of smart cities is becoming increasingly popular. Technology alone, however, is not sufficient for addressing the challenges that modern cities face. Pressures stemming from pollution, stress, and changing environmental conditions can ruin the lives of city dwellers, putting enormous pressure on public finance to address or mitigate consequences of the issues. Application of Nature-based Solutions could address multiple societal and environmental issues common to modern cities. Furthermore, the solutions can immensely benefit from integration with modern technologies. In fact, modern technologies can enhance the implementation, monitoring, and scalability of Nature-based Solutions. This makes a strong case for the application of Nature-based Solutions in modern urban environments that aspire to promote technology and become smart cities. Integration of the solutions into smart cities, however, is not a trivial task and requires a holistic approach to city planning and deep understanding of the ways to do so. Justification of the associated costs requires thorough understanding of the benefits, including the values that may be easily overlooked. Thus, in this study, we apply a conceptual research approach to explore how Nature-based Solutions can be integrated into Smart Cities, what are key benefits of the approach, and how this integration can address significant challenges in urban environments.
Authors - Victor Sineglazov, Andrew Sheruda Abstract - In this paper, a new hybrid segmentation method based on SSL (semi-supervised learning) was developed for samples with image sequences, not all of which were labeled. Thus, this method can find application in areas where labeling is expensive or requires a certain specialist, such as in medicine. The developed method was evaluated on a sample of echocardiography images of patients with infective endocarditis in the context of a real-world task of segmenting heart valve anomalies. As a result, the accuracy gain compared to supervised learning is 5% in the IOU metric, while with other SSL methods it is on average 3%.
Authors - Ervin Mikhail G. Garcia, Zara Naomi S. Inocencio, Ronnel Christian B. Langit, Paul James L. Perez, Harold Russell P. Visperas, Mary Jane C. Samonte Abstract - SHair is an innovative and very useful web-based tool that is made to make the process of donating hair easier. Its primary objective is to aid cancer patients in regaining self-confidence. SHair empowers hair donation by providing a user-friendly site that is safe for anybody wishing to give their hair. The main goal of the website is to have a substantial impact on the well-being of cancer patients by facilitating the connection between persons who express the intention to contribute their hair and those who need facial hair transplants. Individuals must possess this quality to foster a comprehensive comprehension of one another and collectives must possess it to increase their resilience. People stress how easy hair renewal is because it can bring donors and patients together, which is good for patients' mental health. SHair also makes sure that the processing and distribution of given hair can happen, which helps with efforts that focus on the happiness and well-being of cancer patients.
Authors - Denis Vasiliev, Rodney Stevens, Lennart Bornmalm Abstract - Implementation of Nature-based Solutions is becoming increasingly widespread. The solutions are intended to simultaneously address environmental, social and economic challenges. This approach is intended to foster sustainable development. However, mere use of nature for addressing specific problems does not necessarily result in simultaneous delivery of value in all three sustainability areas. To make sure that Nature-based Solutions serve both nature and society and deliver maximal benefits, a holistic approach to their implementation is essential. Implementing this approach is, however, not a trivial goal. It requires joint consideration of environmental, social and economic factors at a range of spatial and temporal scales. Furthermore, collaboration among multiple diverse stakeholders in the context of rapidly changing systems is essential. This often involves processing large volumes of data and can be very labor intensive. As a result, the costs of such projects may be overly high, hindering their implementation. The emerging technology of Generative Artificial Intelligence can greatly facilitate the process, bringing down the costs and increasing speed and feasibility of the project implementation. However, lack of awareness and understanding of how such tools can be used in Nature-based Solutions projects may result in missing these opportunities. Thus, this paper explores potential applications of Generative Artificial Intelligence tools in pro-jects involving Nature-based Solutions.
Authors - Md Asif Ahmed, Md Sadatuzzaman Saagoto, Farhan Mahbub, Protik Barua Abstract - Graphene is emerging as a strong candidate for qubit applications in quantum computing due to its unique properties and recent technological advancements. Graphene, as a two-dimensional material with high carrier mobility and distinct electron behavior, presents potential advantages for qubit applications. However, its zero-band-gap nature poses challenges for stable quantum states, requiring innovative solutions to realize its full potential in quantum computing. This review explores graphene's unique properties and their impact on qubit design, analyzing recent breakthroughs aimed at overcoming its inherent limitations, such as techniques for band-gap modulation and substrate engineering. We delve into various methodologies, including the integration of hexagonal boron nitride (hBN) and electrostatic gating, to enhance graphene's performance for quantum applications. Additionally, we examine the integration of graphene with other 2D materials and hybrid structures to achieve tunable quantum properties, essential for advancing scalable quantum architectures. This comprehensive analysis aims to bridge the material science challenges with the practical demands of qubit technology, providing a roadmap for leveraging graphene in future quantum systems.
Authors - Kobus Kemp, Lynette Drevin, Magda Huisman Abstract - This paper reports on a study that explores and addresses security challenges in the development of enterprise mobile applications (EMAs). Despite the growing prevalence of mobile applications, security considerations are often overlooked or insufficiently addressed in mobile application software development methodologies. This gap highlights the need to incorporate security training into software developer education. The study used a literature review of software development methodologies (SDMs) and security practices, complemented by case studies involving interviews with industry experts on EMA development processes. Using thematic and cross-case analyses, the study produced a framework designed to guide the integration of security measures into EMA development. Findings revealed a limited emphasis on security aspects in current mobile application development practices. Consequently, a partial framework is presented in this paper, detailing key security considerations and countermeasures specific to EMA development. This research contributes to the discipline by offering developers guidelines to enhance security in EMAs, emphasizing the importance of integrating these practices into developer training programs.
Authors - Marisol Roldan-Palacios, Aurelio Lopez-Lopez Abstract - Limited available data becomes a critical problem in specific machine learning tasks, where approaches, such as large language models, turn impractical. Reaching solutions in such situations requires alternative methods, especially whether the object of study contributes to data scarcity while preventing using techniques such as data augmentation. This scenario led us to formulate the research question on how to squeeze hidden information from small data. In this work, we propose a data processing and evaluation technique to increase information extraction from scarce data. Attributes expressed as trajectories are further pair-related by proximity and assessed by customary learning algorithms. The efficacy of the proposed approach is tested and validated in language samples from individuals affected by a brain injury. Direct classification on raw and normalized data from three sets of lexical attributes works as a baseline. Here, we report two learning algorithms out of five explored, showing consistent behavior and demonstrating satisfactory discriminatory capabilities of the approach in most cases, with encouraging percentages of improvement in terms of f1-measure. We are in the process of testing the approach in language data sets of syntactic and fluency features, but other fields can take advantage of the technique.
Authors - Khalaf Elwadya, Khosro Salmani Abstract - The evolution of social media platforms has led to the creation of a dynamic ecosystem, abundant in user-generated content. This, however, has also resulted in raising concerns about data privacy. Beyond potential threats like scammers exploiting freely shared information on social media for spying, financial scams, social media companies can leverage user data to sell targeted advertising. Addressing these issues necessitates heightened user awareness. Hence, this paper first examines the privacy policies of major social media platforms including TikTok, Twitter, Facebook, Instagram, and LinkedIn, providing a comparative analysis of their data storage practices, utilization of user information, account verification requirements, and default privacy settings. Next, we undertake an extensive survey utilizing the data gathered in the initial phase to evaluate user awareness regarding the utilization of their data, highlighting a notable gap between policy stipulations and user expectations. We conclude with four recommendations based on our findings to help social media companies refine their privacy policies, promoting more comprehensible guidelines.
