CAREER: Bringing Structure to the Unstructured: Robust Causal and Statistical Modeling of High-dimensional Unstructured Data

Regents Of The University Of Michigan - Ann Arbor, Ann Arbor MI

Investigators

Abstract

This project focuses on developing new analytical tools for making informed decisions based on complex data types, such as text, images, and gene expressions. Traditional methods of analysis often rely on structured data, but modern challenges require innovative approaches to understand cause-and-effect relationships in high-dimensional datasets. The goal is to improve our ability to extract meaningful insights from diverse sources of information, enabling researchers and practitioners to draw reliable conclusions even when dealing with large amounts of unstructured data. By creating new causal inference methods and training the next generation of data scientists, this project aims to enhance our understanding of complex systems and promote inclusivity in the field of machine learning and data science. This project involves three primary research objectives to advance causal inference for high-dimensional unstructured data. Firstly, we will develop novel causal inference methods specifically designed to accommodate high-dimensional treatments or outcomes, enabling more accurate modeling of complex cause-and-effect relationships. Secondly, we will design probabilistic causal representation learning algorithms to uncover latent causal variables beneath high-dimensional observations. Finally, we will investigate new extrapolation and experimental design techniques for causal inference with unstructured data, allowing us to better design experiments that can effectively tease apart cause-and-effect relationships. Through these research activities, we aim to significantly enhance current methods in machine learning and causal inference, ultimately enabling more reliable and effective causal analysis across diverse scientific fields. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.