UPLB ICS Peak One

This study introduces ”Seated,” a web-based reservation system that aims to enhance the booking process for coworking spaces. The mobile-first application seeks to im- prove user experience and operational efficiency by minimiz- ing miscommunication and delays inherent in current manual systems. Targeting the Banned Books Reading Room in Los Ba˜ nos, the system will feature seat availability visualization, reservation requests, and management tools for customers, staff, and administrator. Using Vue.JS, Vuetify, Express, and Firebase, the platform ensures responsiveness and reliability. Literature reviews reveal a gap in user-friendly reservation systems for coworking spaces, drawing insights from the hospitality and transportation sectors to guide design and implementation. By adopting this digital solution, coworking spaces can better meet rising demand and improve customer satisfaction.

In the Philippines, one of the heaviest workloads that teachers have is the checking of student outputs. Automated Essay Grading (AEG) using Large Language Models (LLMs) is a potential solution that may address the problem but needs to be improved due to high computational or hardware costs. This study investigates the effectiveness of emerging compression techniques, namely Parameter-Efficient Sparse Training (PST), Retraining-Free Pruning (RFP), Generative Post-training Quan- tization (GPTQ), and ZeroQuant on BERT. The uncompressed, finetuned BERT model served as the base model to which the compressed BERT models are compared. The base model was finetuned on the ASAP dataset, composed of general-knowledge short essays with lengths ranging from 150 to 650 words that test writing, reasoning, and expression in the English language. The base model achieved Substantial Agreement using the QWK metric and Moderate Agreement with Krippendorff’s Alpha metric. Additionally, at a 75% compression rate, the model compressed using GPTQ, and the model compressed using PST achieved the same agreement scores on both QWK and Krippendorff’s Alpha as the base model.

Computer graphics is a pillar of modern computing, enabling the generation of digital shapes, displaying of images and videos, creation of graphical user interfaces, video game development, and much more. Courses that focus on teaching foundational computer graphics lessons traditionally use a web- based approach, with the Open Graphics Library (OpenGL) and the Web Graphics Library (WebGL) for instruction. This study aims to supplement the delivery of computer graph- ics lessons using an augmented reality (AR) application, AR Graphix. Developed with SwiftUI, UIKit, ARKit, and SceneKit, the application offers modules to teach key concepts in computer graphics and allows the user to upload and use their own 3D models. Preliminary testing with 13 respondents yielded a System Usability Scale (SUS) score of 90.77.

This study developed and evaluated a mobile ap- plication designed to recommend recipes based on users’ food preferences, allergies and nutritional needs, and available ingre- dients. Utilizing technologies such as Flutter, Firebase, Cloud Firestore, and APIs from Edamam and Spoonacular, the app provides personalized recipe recommendations and substitutes for ingredients identified as allergens. Usability was assessed using the System Usability Scale (SUS), which yielded a high mean score of 91.07, obtained from 28 students. This indicate very good user satisfaction and ease of use. User feedback highlighted the app’s strengths, such as its comprehensive search feature and intuitive interface, while also identifying areas for improvement. The application successfully meets its objectives, demonstrating significant potential as a valuable tool for personalized culinary guidance and modern dietary needs.

This study introduces DeskUP, a web- application developed for the University of the Philippines - Los Ba˜ nos (UPLB) to address the challenges faced by administrative staff in managing inquiries from students. With UPLB’s complex administrative structure and limited staff resources, DeskUP aims to streamline the inquiry process by providing a centralized platform for submitting, managing, and resolving queries. Through user authentica- tion, ticket management, and office-specific functionalities, DeskUP enhances efficiency and user experience, catering to both students and administrative staff. It garnered positive usability ratings from student respondents (SUS score of 78.14) and less favorable feedback from staff respondents (SUS score of 66), indicating areas for im- provement to further optimize system performance and user satisfaction.