Projects

Online Throughput SchedulingJul 2025 - Present

Topics: Online Algorithms, Competitive Analysis

This individual research investigates online throughput scheduling problem under various preemption models. Currently exploring the results for randomized algorithms for the preemption-revoke model. Additionally, I consider the discrete-time perspective, where all the parameters of jobs are integers and the scheduler makes decisions at discrete time steps.

Supervised by Professor Allan Borodin at the University of Toronto.

A. Borodin, C. He, N. Mottu. arXiv preprint (2026). arXiv: 2607.16163.

C. He. arXiv preprint (2025). arXiv:2510.15318.

Poster presented at the CSSU Undergraduate Research Conference, University of Toronto (2026). [View Poster]

Nonogram Puzzle Difficulty Analysis Sep 2025 - Aug 2026

Topics: Human-Perceived Difficulty, SAT Solvers

This research investigates how computational measures of difficulty align with human perceptions in Nonogram puzzles. We formulate Nonograms as Boolean satisfiability (SAT) problems and analyze solver-derived metrics such as decisions, propagations, and conflicts. Using these metrics, we generate and select puzzles across a range of computational difficulty levels.

To evaluate human difficulty, we conduct a web-based human-subject study measuring solving time, actions, hints, and subjective difficulty ratings. By comparing solver-based metrics with human performance, this work aims to identify computational indicators that better reflect human puzzle difficulty and support the generation of human-aligned Nonogram puzzles.

Co-supervised by Professor Alice Gao and Professor Jonathan Calver at the University of Toronto.

C. He, Y. Ju., A. Gao, J. Calver. Submitted to AAAI 27; arXiv preprint (2026). arXiv: 2608.23300.

Poster presented at the CSSU Undergraduate Research Conference, University of Toronto (2026). [View Poster]

Monte Carlo Tree Search for Othello May 2025 - Aug 2025

Topics: Artificial Intelligence, Search Algorithms

In the second phase of the project, I focused on advanced AI techniques for strategic gameplay in Othello. I implemented multiple Monte Carlo Tree Search (MCTS) agents, exploring various configurations including random and heuristic-based rollout policies, partial and full expansion strategies, and tuning of exploration constants. I conducted large-scale performance evaluations to study how these design choices impact win rates and efficiency.

Supervised by Professor Alice Gao at the University of Toronto.

Poster presented at the Undergraduate Summer Research Showcase, University of Toronto (2025). [View Poster]

Designing and Implementing AI Agents for Othello Jan 2025 - Apr 2025

Topics: Artificial Intelligence, Game AI

I designed and implemented a fully functional Othello (Reversi) game engine in Python, supporting human input, random agents, and AI agents. This included developing the full game logic—legal move generation, disc flipping, and end-game detection—within a modular architecture. I built several game-playing agents, including a minimax search agent enhanced with alpha-beta pruning and various heuristic functions such as coin parity, mobility, stability, and corner control. The system was equipped with a flexible command-line interface and supported automated batch evaluations between agents for performance testing under different settings.

Supervised by Professor Alice Gao at the University of Toronto.

Poster presented at the Undergraduate Summer Research Showcase, University of Toronto (2025). [View Poster]