AI in Science Fellow

In data from telescopes, exoplanet signals are masked by complex instrumental and astrophysical noise sources of apriori unknown form. Conventional exoplanet-detection pipelines perform sequential data processing to mitigate noise, relying on incomplete signal models at each step, which can lower overall detection performance. My research analyzes the statistical properties of exoplanet data-processing architectures to design ...

Freshwater ecosystems are facing an escalating water quality degradation globally. During my PhD at Georgia Tech, I focused on leveraging process-based models to study complex water quality/ecology dynamics. With the burgeoning availability of environmental data and computational resources, I have recognized the potential of AI techniques, particularly for large-scale studies. Therefore, as a Schmidt AI ...

Xinyu (Cindy) Liu is a final-year Ph.D. candidate in Operations Research at Georgia Tech, with a doctoral minor in Transportation Systems and Engineering. Prior to that, she graduated summa cum laude in Business Analytics from Singapore University of Technology and Design. Xinyu’s research interest is to develop stochastic optimization and models, data-driven methods, and computational tools to ...

Biocatalysis bears great potential to accelerate the synthesis of functional organic molecules by leveraging the ability of enzymes to catalyze reactions remarkably selectively. To facilitate its application, we aim to predict suitable enzymes for new substrates by viewing biocatalysis as a heterogeneous enzyme-substrate network. The models will be prospectively applied toward preparing pharmaceutically relevant molecules ...

I develop statistical methods and theory to study the effects of natural selection and other evolutionary processes on observed DNA sequences. Using machine learning techniques, I plan to approximate high-dimensional genetic relationships between individuals. Using generative modeling, I plan to generate artificial chromosomes that reflect complex patterns indistinguishable from those seen in real data, particularly ...

My project aims to significantly improve debris detection capability to the size range of lethal non-trackable space debris (10 μm – 10 mm) using ground-based instruments to detect non-thermal electromagnetic (NTEM) radiation. With the help of AI/ML models, detecting debris signatures from DSN data can be done automatically and efficiently. Science Mentor: Mojtaba Akhavan-Tafti, Climate ...

Madeline is a proud Pittsburgher and Torontonian with training in mathematical and computational biology, particularly the ecology and evolution of malaria. Madeline’s current research interests focus on developing our quantitative understanding of basic biological processes that underly within-host infection dynamics. She is working towards a new approach for developing compact, predictive models of within-host infection ...

Xin’s research interests include risk, reliability, and resilience analysis of natural hazards under climate change, leveraging interdisciplinary approaches such as remote sensing, geospatial modeling, data-driven and physics-based models. Before joining the University of Michigan, he earned his PhD in Civil Engineering from Shanghai Jiao Tong University under the supervision of Prof. Lulu Zhang. He was ...