NJ / FengDian Su

This study enhances cross-lingual information retrieval by applying LLM reranking, improving performance by 82% compared to the original methods, and utilizes retrieval-augmented generation for more accurate responses.

We developed a light novel-focused database management system using SQLite and Flask. This system provides users with an efficient and user-friendly platform for managing and accessing light novel data.

We implemented an intelligent power monitoring system for users' homes using AIoT technology. Providing real-time electricity data and cost forecasts on a web interface through power measurements and backend data analysis.

Based on Shamir's secret sharing technique, we implemented a feature that hides plaintext within a QR code. Users must gather a specified number of participants to retrieve the actual message.