Haowei Liu

3rd-year Ph.D. Student in Computer Engineering, with research focused on large language models, text ranking, and information retrieval. Experienced in building and evaluating machine learning systems, including fine-tuning, PEFT methods, and experiment pipelines. Skilled in Python, PyTorch, and HuggingFace tools, with hands-on work in dataset construction, model analysis, and reproducible research workflows. Looking for a Ph.D. Research intern role in machine learning or natural language processing.

Wei, T. R., Liu, H., Wu, X., Fang, Y.
A survey on feedback-based multi-step reasoning for large language models on mathematics. arXiv:2502.14333, 2025.

Liu, H., Wu, X., Sun, G., Tao, Z., Fang, Y.
RaCT: Ranking-aware Chain-of-Thought Optimization for LLMs. In SIGIR-AP 2025.

Wei, T. R., Liu, H., Hu, H. C., Wu, X., Fang, Y., Wu, H. T.
CLERF: Contrastive LEaRning for Full Range Head Pose Estimation. arXiv:2412.02066, 2024.

Hu, H. C., Wu, X., Liu, H., Wei, T. R., Wu, H. T.
Full-range Head Pose Geometric Data Augmentations. arXiv:2408.01566, 2024.

NPC Decision Making with LLMs

• Built an LLM-driven NPC decision system that produces context-aware actions and dialogue in game environments.
• Integrated reinforcement learning, retrieval-based methods, and fine-tuned models to support consistent NPC behavior.
• Designed interaction loops that allow NPCs to respond to player actions in real time.

Knowledge Editing for Large Language Models

• Developed methods to update specific model facts without full retraining using targeted fine-tuning and memory-efficient techniques.
• Evaluated multiple intervention-based approaches to study how updates affect model performance.
• Designed experiments to measure stability, retention, and side effects of knowledge edits.