Research Focus
Diffusion models have become a popular machine learning approach for many types of problems, such as linear inverse problems. Paleoclimate reconstruction based on climate proxies such as tree rings and pollen is an inverse problem. Many statistical methods have been used to reconstruct past climates, and we plan to investigate how diffusion models can be leveraged to improve performance compared to existing statistical methods.
Skills, Techniques, Methods
- Computing
- Python
- Basic understanding of statistics
Research Conditions
The student should expect to read machine learning literature and try implementing algorithms. They are expected to use their own computer for research.
Team Structure and Opportunities
Undergraduates will meet regularly with Dr. Li and her collaborators. Students are expected to present progress at each meeting.
Requirements
Basic skills in computing software like R or Python, and at least an introductory level of statistics. Students with more advanced statistical and/or machine learning skills are preferred.
