Authors - Al John A. Villareal, Jaime M. Samaniego Abstract - Potholes significantly impact road safety, vehicle performance, and infrastructure maintenance, particularly in developing countries where monitoring systems remain largely manual. This study presents the design and implementation of an Android-based mobile application that utilizes sensor fusion and deep learning for real-time pothole detection and severity analysis. The system integrates a YOLO-based object detection model with smartphone sensors, including accelerometer, gyroscope, and Global Positioning System (GPS), enabling simultaneous visual and motion-based detection. A dataset consisting of 9,253 road surface images containing 16,123 pothole annotations was used for training and evaluation using a 70:20:10 dataset split for training, validation, and testing. Among the evaluated models, YOLO11s achieved the highest mAP@50– 95 value of 54.2%. However, YOLO26n was selected and implemented in the developed Android application due to its competitive detection performance, compact 5.2 MB model size, and suitability for real-time mobile deployment. Field testing across four road segments covering 18.97 kilometers resulted in 130 detections, of which 84 were verified potholes and 46 were false detections, yielding a verification rate of 64.62% and a false detection rate of 35.38%. The system recorded an average detection density of 6.85 potholes per kilometer. Results demonstrate that integrating deep learning and sensor fusion in a mobile platform provides a scalable and cost-effective solution for automated road condition monitoring and intelligent transportation systems.
Authors - Nu Yin Khaing, Win Lelt Lelt Phyu Abstract - Water quality monitoring is essential for sustainable aquaculture management and fish health assessment. However, monitoring a large number of physicochemical parameters in creases sensor deployment costs and system complexity. While traditional machine learning ap proaches struggle with complex, nonlinear relationships among water quality variables, feature reduction can optimize system efficiency. This study proposes a hybrid machine learning and deep learning framework to achieve accurate, cost-effective water quality classification using the Water Quality Dataset (WQD). The framework integrates Random Forest (RF) and XGBoost for feature selection with Support Vector Machine (SVM) and Long Short-Term Memory (LSTM) models as classifiers, evaluating four hybrid combinations (XGBoost+SVM, XGBoost+LSTM, RF+SVM, and RF+LSTM) across subsets of 10, 7, and 5 features. Experimental results demon strate that hybrid deep learning architectures consistently outperform traditional machine learn ing methods. Specifically, XGBoost+LSTM and RF+LSTM achieved the highest classification accuracy of 96.05% using 10 selected features, while maintaining reliable performance at lower dimensions. Furthermore, Shapley Additive explanations (SHAP) analysis was applied to en hance model interpretability, identifying Dissolved Oxygen (DO), Turbidity, BOD, H2S, and Nitrite as the most important attributes. Ultimately, the proposed framework minimizes sensor requirements and provides an accurate, interpretable, and economically viable solution for aqua culture water quality monitoring systems.
Authors - Apolinar P. Datu, Barnard J. Maraon, Rommel H. Orquiza, Cristopher T. Takano, Olivia L. Yosa, Mark Joseph G. Cruz, Jeferson D. Talisayon Abstract - This study examines the integration of multilingual customer service skills into workforce development programs for hospitality management students. In today’s globalized environment, the hospitality industry serves guests from diverse cultural and linguistic backgrounds, making effective communication an essential skill. Using a quantitative approach, the study gathered data from 100 hospitality students to assess their communication skills, performance, and confidence in multilingual settings. The findings reveal that most students already possess basic multilingual abilities, particularly in English and Filipino, which serve as a strong foundation for further development. Results also show that the current curriculum includes elements of multilingual training, contributing to students’ overall competence. However, while students demonstrate satisfactory performance and confidence, there is still a need for increased practical exposure and strengthened training programs. Furthermore, the study highlights that students generally meet industry expectations but may benefit from more structured and continuous training to enhance their real-world readiness. Overall, integrating multilingual customer service skills significantly supports the development of hospitality students, preparing them for the demands of a culturally diverse workforce and improving their competitiveness in the global hospitality industry.
Authors - Sanjana Priyadarsini, Choudhary Aman Kumar Roy, Ashlesha Shree Bajpai, Rajdeep Banerjee, Shivali Sharma, Ranjita Kumari Dash Abstract - Today, machine learning methods are quickly being adopted in healthcare. In numerous instances, it has been observed that datadriven approaches have increased reliance of medical data analysis and disease detection by about 60-70%. It is important to diagnose cardiac arrhythmias early using electrocardiogram (ECG) analysis, as timely diagnosis can prevent severe complications and loss of life. Most ECG datasets are however not balanced, with normal beats by far outnumbering abnormal ones and causing the models to underperform on rare but significant cases. In this work, Logistic Regression is used as a baseline model. To correct this imbalance, Class weighting and Synthetic Minority over-sampling Technique (SMOTE) are applied. These techniques help the model detect rare heartbeat patterns more reliably and miss fewer abnormalities. This paper shows that addressing class imbalance can make ECG-based classification systems more accurate and clinically valid.
Authors - Windy Permata Suyono, Dwi Handarini, Eka Septariana Puspa, Surya Anugrah, Nuramalia Hasanah, Ratna Anggraini, Sabo Hermawan, Rio Firnanda Abstract - This study aims to develop an AI-Based Digital Audit Maturity Frame work integrated with SPIP to support Smart SME Governance and Sustainable Development Goal 9 (SDG 9). The study employs a systematic literature review approach by analyzing 50 relevant articles published between 2020 and 2026 re lated to artificial intelligence in auditing, digital audit maturity, internal control systems, smart governance, SMEs, and sustainable innovation. The findings in dicate that artificial intelligence technologies significantly improve audit effec tiveness, governance transparency, operational efficiency, and organizational re silience. The study also reveals that digital audit maturity and SPIP-based internal control systems play important roles in supporting sustainable digital transfor mation and adaptive governance within SMEs. Based on the literature synthesis, this study proposes a conceptual framework consisting of input factors, digital transformation processes, digital audit maturity levels, SPIP-based internal con trol systems, smart SME governance, and SDG 9 achievement. The proposed framework contributes theoretically by integrating technological capability, gov ernance systems, internal control mechanisms, and sustainability perspectives into a unified governance model. Practically, the framework provides guidance for SMEs, policymakers, auditors, and digital transformation practitioners in strengthening sustainable AI-driven governance implementation.
Authors - Milind Nemade, Khush Chheda, Rahul Dhanak, Durgeshkumar Dubey Abstract - The problem of meeting productivity continues to prevail in the current era in multilingual environments due to frequent language switching between speakers. Most of the existing frameworks for meeting intelligence primarily focus on automatic transcription and lack significant support for Indic languages, speaker identification, and task extraction. Additionally, many of these frameworks depend on metadata associated with specific platforms and, therefore, cannot be used in any offline environment or even on other platforms. In this work, we present a scalable and platform-agnostic framework for meeting intelligence which can automatically an alyze meetings post factors by leveraging speech recognition, speaker identification, and contex tual analysis. The system leverages multilingual and code-switched transcription capabilities of Sarvam AI, generates speaker embeddings using ECAPA-TDNN, and then uses Large Language Models for context-based analysis. Two different strategies for speaker identification are dis cussed in this paper such that they do not need any platform-based metadata while improving the attribution accuracy. We have further developed an asynchronous framework for extracting tasks, assigning tasks, and notifying about them. Experimental results indicate enhanced transcription accuracy as well as speaker identification accuracy in Hindi-English code-switching cases. Fu ture work will focus on implementing advanced privacy protection and end-to-end encryption mechanisms for secure storage and processing of meeting recordings and metadata.
