Our mission is to develop the next generation of advanced machine learning tools for controlling complex physical systems by discovering physically interpretable and physics-constrained data-driven models through optimal sensor selection and placement. Our work is anchored by a common task framework that evaluates the performance of machine learning algorithms, architectures, and optimization schemes for the diverse tasks required in engineering applications. We will push beyond the boundaries of modern techniques by closing the loop between data collection, control, and modeling, creating a unique and cross-disciplinary architecture for learning physically interpretable and physics constrained models of complex dynamic systems from time series data. The common task framework will further support sustainable and open-source challenge datasets, which will be foundational for developing interpretable, ethical, and inclusive tools to solve problems fundamental to human safety, society, and the environment.

Founded Oct. 1, 2021, as part of the National Science Foundation's effort to advance machine learning and AI across the sciences, the Institute is committed to integrating machine learning and artificial intelligence methods for a broad range of scientific and engineering applications. The $20 million investment integrates institutions from the Pacific Northwest (Washington, Montana State, Nevada, Hawaii, Portland State, Boise State and Alaska) with Harvard and Columbia Universities. Our aim is to not only develop fundamental ML/AI methods and algorithms, but also to broadly share these methods with the science and engineering communities by providing open-source code, data, and lectures on the broad and diverse range of topics we consider as an Institute. Thus, we have three key efforts: fundamentals, applications, and education.