Computer Sciences professor and department chair Steve Wright has been elected to the 2024 class of the National Academy of Engineering (NAE) in recognition of his achievements in the design and theory of optimization algorithms and the application of optimization to such areas as machine learning and signal processing.
NAE membership honors those who have made outstanding contributions to “engineering research, practice, or education, including, where appropriate, significant contributions to the engineering literature.”
“I was shocked but delighted to receive the news by email, just two hours before it was publicly announced,” Wright said. “I had to read the message several times to make sure it was real.”
Wright’s chief research area of optimization is mathematical and computational at its core but can be applied to an enormous range of applications that involve maximization or minimization of some quantity. As examples: An investor may want to allocate their money among different stocks in a way that maximizes their expected return while limiting some measure of risk, or the efficiency of a chemical production process can be optimized by choosing the best combination of flows, temperatures, and pressures in the plant. In machine learning, a neural network can be trained to recognize objects in a picture by showing it an enormous number of pictures and adjusting its parameters so that it identifies the objects in all these pictures correctly. In practice, these adjustments are made little by little, so as to minimize the errors made by the neural network in identifying the objects in the images – an optimization problem! In fact, the centrality of optimization to the AI revolution has given rise to a new ferment of research in optimization and has drawn many new researchers and practitioners to the field.
Much of Wright’s work concerns the design of optimization algorithms with strong mathematical properties for solving standard optimization paradigms that can be applied across a wide range of fields. He has focused on methods for optimization problems over continuous variables, an area covered in his highly influential book Numerical Optimization, co-authored with J. Nocedal. Besides being a popular textbook, it became a standard reference for researchers in machine learning once the AI revolution took flight. Wright contributed to a surge of interest in interior-point methods (for linear programming, a canonical optimization problem) during the 1990s, summarizing the outburst of research in this area in a 1997 book that became a touchstone reference. During the explosion of interest in compressed sensing, a revolutionary technique in signal processing that became popular from 2005 onward, Wright and his collaborators developed the optimization algorithms that are crucial for solving problems in the field. These same techniques, and others closely related, turned out to be crucial in machine learning too.
The NAE was founded in 1964 and is part of the National Academies of Sciences, Engineering, and Medicine. The NAE’s mission is “to advance the welfare and prosperity of the nation by providing independent advice on matters involving engineering and technology and by promoting a vibrant engineering profession and public appreciation of engineering.”
A UW-Madison faculty member since 2001, Wright’s honors and awards include the Khachiyan Prize in 2020 from the INFORMS Optimization Society (for lifetime achievements in optimization) and three named chairs/professorships at UW-Madison: the George B. Dantzig Professorship, the Sheldon B. Lubar Chair, and the Amar and Balinder Sohi Professorship. He became a Fellow of SIAM in 2011 and received the Hilldale Award from UW-Madison in 2022.