UPLB ICS Peak One

The exponential surge in malware types, from 100 million in 2012 to 700 million in 2018, poses a significant threat to the cybersecurity sector. While deep learning models have demonstrated impressive accuracy in malware detection and classification, even exceeding 99% in recent studies, the underlying features driving these predictions remain elusive. This lack of transparency hinders user trust in the models’ decisions. To address this, this research explores the use of XAI tools to analyze the inner workings of black-box malware detection models trained on tabular and NLP-based data from the EMBER dataset. By uncovering the influential features in single data points, this study aims to enhance model transparency, foster trust in black-box malware detection predictions, and pave the way for improved detection methods and more effective threat analysis in the ever-evolving landscape of cybersecurity.

The MaizeField Registration and Data Manage- ment System is a web-based platform designed to streamline registration, profile management, and ID generation for corn farmers under the Philippine Maize Federation Inc. (PhilMaize). Featuring integrated QR codes for verification and a discount system framework, it enhances efficiency and accuracy in data handling. The system also incorporates robust security measures to protect sensitive farmer data and ensure system reliability. Deployed during the 17th Philippine National Corn Congress, the system successfully facilitated farmer registrations, profile management, and event coordination, reducing manual effort and improving data reliability. This project demonstrates how digital solutions can modernize agricultural operations and lays the foundation for further innovations to support the farming community.

Campus events offer opportunities for student growth through self-discovery, skill development, and social engagement. However, the absence of a dedicated platform to access event information limits the impact of these events. UPLB Connect was developed to address this gap, particularly for the benefit of the constituents of the University of the Philippines Los Ba˜ nos (UPLB). This mobile application provides a centralized and streamlined events feed that consolidates organization events and official university events from the official academic year calendar into a single chronological newsfeed for efficient and comprehensive event discovery. A reminder system was also implemented to keep students updated on events that matter to them. It offers organizations a focused platform for better event publicity. Additionally, UPLB constituents could benefit from the data visualizations on the dashboard to identify trends and draw insights. The application features distinct interfaces for regular users, organizations, and admins. With an overall System Usability Scale (SUS) score of 85.139 from the evaluation of 18 respondents, UPLB Connect demonstrated an acceptable, ”Best imaginable” level of usability and user-friendliness. Moreover, the positive feedback from respondents, who explicitly appreciated key features of the application, further emphasizes the platform’s relevance and reflects the fulfillment of the study’s objectives.

Mariz’ sari-sari store and Tony’s meat shop are mi- cro enterprises that experienced decreasing sales due to increased competition in the area after the COVID-19 pandemic. They had to digitize through a mobile-based online shop. The application provides customers with a user-friendly interface to browse and order essential goods, and pay using cash and online payment methods. They can also pay later like traditional local stores, and manage their orders. Moreover, it also provides business owners with inventory, user, and order management to assist them in facilitating the stores. 23 respondents evaluated the system’s usability, learnability, and information and interface quality using the Post-Study System Usability Questionnaire (PSSUQ). Results show an excellent score of 1.332 and feedback that provides insight into users’ needs.

Remote learning combines online and offline edu- cational resources to enhance accessibility. In the Philippines, remote learning faces problems due to inconsistent Internet connectivity, thus needing to develop offline solutions. This paper aimed to address this challenge by developing an application to enable educators to generate offline course packages deployed as a client-side only web application for students. The application demonstrated interoperability across Windows, macOs and Linux operating systems. The application was evaluated by educators from University of the Philippines Los Banos (UPLB) and University of the Philippines Open University (UPOU) using the Systems Usability Scale (SUS) resulting in a score of 80.00 which is considered good. A sample offline course package was evaluated by students in various University of the Philippines (UP) campuses using the SUS resulting in a score of 85.00 which is also considered good. The application can help support diverse educational needs and contribute to the broader implementation of Massive Open Online Courses (MOOCs) in remote regions.