Hi, I’m Theresa! I’m a 5th year Computer Science PhD student at the University of Minnesota working in the Knowledge Computing Lab with Professor Yao-Yi Chiang.

Download my CV here.

Research Interests

• Machine learning for spatial problems and geographic data. Specifically, my background is in computer vision and state-of-the-art vision models to better address challenges unique to environmental data.

• Fast adaptation of geo-foundation models for multimodal environmental data.

• Uncertainty quantification across geographic space in deep neural networks. Specifically increasing interpretability by combining spatial statistics and uncertainty modeling techniques in the deep learning domain.

Projects

Ongoing

• CEDAR: Carbon Estimation with Deep Learning (part of the AI-LEAF initiative)
• Peatlands Permafrost Mapping (part of the AI-LEAF initiative)

Previous

• Automatically Georeferencing Geologic Maps
• Post-disaster Building Damage Assessment
• Machine Learning for Species Distribution Modeling
  • Developed a transformer-based model to learn geographic embeddings from multimodal, multi-resolution data in order to ultimately predict the distribution of bird species across the United States.
• Using Deep Neural Networks to Generate Representations of Urban Neighborhoods
  • Understanding urban environments by generating indicators, such as walkability or greenness, requires us to understand how the city is split up; i.e. what the neighborhoods are. By utilizing street view images and publicly available data, such as the human settlement layer, we automatically clustered urban areas into likely neighborhoods using self-supervised machine learning methods.