10th International Conference on ICT for Intelligent Systems in New York

Authors - Ernesto Rivera-Alvarado, Saul Guadamuz
Abstract - Predatory journals are characterized by publishing research articles without the proper peer-review process. They are mainly focused on getting money from authors, regardless of the quality of the article being submitted, in a scenario where academics and researchers are usually incentivized to create a high volume of research to advance their career paths. While publishing in a predatory journal is easier than in a reputable publisher, the author’s credibility could be tarnished. The rise of large language models has provided a new danger for this scenario. It is now easier to create low-quality research papers using IA tools that could be sent to predatory publishers that fail to perform a rigorous peer-review process. A bad combination of fast and poorly created research papers with highly available pay-to-publish journals provides an unsafe scenario for science publishing and unaware scientists. This paper aims to warn researchers about the dangers of generating low-quality research and publishing in predatory journals.

Authors - Mohammadmahdi Ghobadi, Renee Bryce
Abstract - Modern applications are built from many small services that work together. In these systems, an API gateway plays a key role in routing requests, balancing the load, and processing events. Most API gateways use fixed rules that do not change with shifting workloads, causing delays as high-priority events may be blocked by less critical tasks. In this paper, we propose a new method that uses reinforcement learning to create adaptive API gateways. Our system learns from real-time performance data such as CPU usage, memory usage, event latency, arrival rates of high-priority events, and throughput of low-priority events. By using multiple metrics, the system adjusts its settings to respond to changing conditions. It processes high-priority events immediately while grouping low-priority events into batches for efficient processing. We use the Proximal Policy Optimization (PPO) algorithm because it is stable and effective for learning the best settings. We evaluated our method using simulations that mimic real-world conditions. The results show that our approach reduces the delay for highpriority events by about 41% and significantly lowers the delay for lowpriority events. The system also uses fewer resources than rule-based gateways. These improvements demonstrate that considering multiple performance metrics can lead to smarter, more adaptive API gateways. Our approach adapts very quickly to varying workloads, ensuring remarkably reliable operation during sudden spikes. Our work shows that reinforcement learning can improve API gateway performance for modern applications. This is important for fields such as finance, healthcare, and emergency systems, where fast and reliable responses are critical.

Authors - Paula Poveda-Sotomayor, Henry N. Roa, Edison Loza-Aguirre, Javier Guana-Moya, Nelson Salgado-Reyes
Abstract - Water treatment facilities require efficient and continuous monitoring to ensure safe drinking water. Traditional Supervisory Control and Data Acquisition (SCADA) systems often face limitations in scalability, cost-effectiveness, and real-time data accessibility, prompting interest in alternative technologies such as the Internet of Things (IoT). This study proposes a cloud-integrated IoT framework explicitly designed for real-time monitoring of key water quality parameters, including pH, turbidity, and flow rate. The framework securely transmits data using MQTT protocol to AWS IoT Core, leveraging AWS Timestream and Grafana dashboards for advanced analytics, visualization, and proactive anomaly detection. The system's feasibility and performance were validated through simulation, guided by the Design Science Research (DSR) methodology. Results indicate significant potential benefits, including improved real-time responsiveness, reduced manual monitoring requirements and enhanced operational decision-making. This framework provides a scalable and secure solution that addresses existing SCADA limitations, laying the groundwork for future integration into smart water management practices.

Authors - S. M. Mizanoor Rahman
Abstract - We proposed and investigated a bidirectional trust-triggered cyber-physical-human (CPH) system framework for human-robot collaborative assembly in flexible manufacturing. For this purpose, we developed a one human-one robot hybrid cell where the human and the robot collaborated with each other to perform the assembly operation of different manufacturing components in a flexible manufacturing setup. In the proposed framework, we configured the human-robot collaborative system in three interconnected components of a CPH system: cyber system (software system), physical system (the robot, sensors and necessary hardware), and human system (the human co-worker, supervisor and work environment). We divided the functions of the CPH framework into three interconnected modules: computing or computation, communication and control. We used a model to compute the human and robot’s bidirectional trust in real-time to monitor the performance of the CPH framework. We evaluated the performance of the CPH framework implementing it on an actual human-robot collaborative assembly setup considering systematic variations in the complexity levels of the computing module. The overall results revealed that variations in computing complexity significantly impacted the performance of the CPH framework in terms of human-robot interactions and task efficiency and quality. The results thus proved that it became easy and effective to monitor the performance and interactional effectiveness of a human-robot collaborative system modularly when the system was configured in the form of a CPH system. The results can transform the design, development, analysis and control of human-robot collaborative systems for various applications such as industrial manufacturing and assembly, military operations, transportation, healthcare and rehabilitation, construction, social services, etc.

Authors - Oleksandr M. Khimich, Elena A. Nikolaevskaya, Pavlo S. Yershov
Abstract - Virtual assistants (VAs) are becoming an increasingly popular tool that helps users find information faster and more accurately. This paper proposes a new AI-powered Virtual Travel Guide (GuideAI) system. The concept, architecture, and implementation of a microservice for intelligent travel search are explored in detail. The study focuses on technical aspects, particularly the integration of GPT for analysis and personalization. It highlights how geolocation and personalization can be combined to create more adaptive, user-responsive systems. The proposed approach enhances travel experiences by providing real-time, tailored recommendations based on user preferences and location. The scientific contribution of this research lies in the development of a novel AI-driven approach to travel assistance, demonstrating how advanced natural language processing and contextual data analysis can improve recommendation accuracy. By leveraging machine learning techniques, GuideAI refines search results, enhances decision-making processes, and optimizes user interactions. This work contributes to the field of intelligent information retrieval and personalized AI systems, offering insights for future applications in tourism and beyond.

Authors - Alvaro A. Casanova, Elvis E. Gaona, Oscar Acosta
Abstract - This paper introduces a stochastic modeling framework for analyzing binaural urban noise in Bogotá, Colombia. A total of 4,200 five-minute audio recordings were collected using binaural sensors placed at seven urban locations over 47 days. To analyze the data, we utilized cloud-based tools, including Inferencer app for machine-learning-based sound classification and Soundmetrics app for extracting acoustic parameters such as LEQ (equivalent continuous sound level), IACC (inter-aural cross-correlation coefficient), and WIACC (weighted inter-aural cross-correlation coefficient). The probabilistic analysis revealed instances of extreme noise events exceeding 85 dB. We identified correlation patterns that informed the development of a predictive model based on discrete-time Markov chains. Which incorporated both noise intensity and sound source classification into composite acoustic states. The simulated trajectories effectively captured the temporal dynamics of urban acoustic conditions, successfully meeting the study’s objective of modeling future noise dynamics. This approach demonstrates a scalable solution for real-time monitoring, statistical characterization, and predictive modeling of complex urban soundscapes, providing actionable insights that support data-driven decision-making in smart city planning.

Authors - Ramiro Ramirez, Yasmany Garcia-Ramirez, Jandry Jaramillo
Abstract - This research explores the potential of mobile applications in improving road safety education and driver behavior. Road traffic crashes, particularly those involving vulnerable road users, continue to represent a significant global public health issue. Traditional road safety education methods often lack engagement and immediate feedback, limiting their effectiveness in promoting long-term behavioral change. This study introduces SafeJourney, a mobile application developed to enhance road safety learning through real-time feedback, location-based alerts, and personalized progress tracking. The application integrates GPS technology and mobile sensors to monitor driving, cycling, and pedestrian behavior, providing users with timely feedback to reinforce safe practices. The development of SafeJourney followed a modified SCRUM methodology, ensuring iterative and efficient progress. Key features include real-time alerts, integration with traffic law databases, and a user-friendly interface designed with Figma and React Native. Testing and validation confirmed the application's functionality, including low-latency tracking, device compatibility, and seamless integration with a cloud-based database. SafeJourney represents a promising innovation in road safety education, offering an interactive and scalable solution to reduce traffic-related injuries and fatalities.

Authors - Gery Prasetyo, Tanty Oktavia, Mohammad Ichsan
Abstract - This research presents the current state of development and implementation of AI predictive analytics for project management support through the lens of Indonesian financial projects, highlighting a practical implementation gap and its impacts towards the fulfilment of enhancements in the core project management aspects of risk prediction, budget and resource estimation, deliverables quality monitoring, and automation of project financial processes. A qualitative methodology is applied to explore the various field experiences and sentiments of five financial project professionals in individual financial firms through direct semi-structured interviews. Analysis of the findings affirms the implementation of machine learning predictive analytics in the four key project areas, with tasks in risk strategy implementation, budget estimation, and resource allocation management revealed as the prominent applications utilized by respondent financial companies. Benefits in streamlining tasks and improvements towards project efficiency are noted by the respondents, along with enhancing team awareness against unseen risks and resource issues, setting new standards for quality management, and reducing human errors in rapid financial processing. The findings further uncovered key barriers in predictive analytics implementation success stemming from system maturity and accuracy issues, and moreover the influence of intangible human factors towards the financial sector business environment and practices. Improving the synergy of the technical capabilities predictive AI systems with human subjectivity and intuition is paramount to ensure successful implementations of predictive project support tools in the Indonesian finance sector.