Undergraduate honors research / Sep 2024 – Apr 2025
Undergraduate Honors Thesis
PNRL: Reinforcement Learning Development for an Undergraduate AI Course
A client-server reinforcement-learning teaching environment with inspectable Q-learning telemetry, extended by a Python–Java bridge for the Game of Amazons course infrastructure.
The University of British Columbia
Supervised by Dr. Yong Gao
10,000+training episodes per session
15+teams using Python–Java bridge
50%development-time reduction
One research project, two connected systems
PNRL exposes the state/action/reward loop, Q-value updates, exploration, and training termination choices inside the course’s Java and SmartFoxServer environment. The related Game of Amazons work uses Py4J so Python AI logic can participate while Java retains the server-facing protocol.
Research contribution
- Engineered a modular client-server Java-based RL system capable of processing 10,000+ training episodes per session with efficient Q-table updates, enabling students to conduct real-time experimentation of RL strategies.
- Implemented an epsilon-greedy Q-learning policy, increasing final cumulative rewards of the RL agent by 15% within the first 100 training episodes.
- Built a Python-Java bridge via Py4J, enabling 15+ teams to use Python ML libraries in a Java game—cutting development time by 50%.
Associated repositories
Two connected implementations from the same honors research effort.