Clayton Scott is a Professor of Electrical Engineering and Computer Science (EECS) at the University of Michigan, with a courtesy appointment in Statistics. He holds a joint appointment in the College of Engineering and is affiliated with MIDAS, AI Lab, and the Center for Computational Medicine and Bioimaging (CCMB). His research focuses on statistical machine learning theory and algorithms, with applications in medical imaging, nuclear engineering, climate science, and clinical diagnostics. He actively collaborates with researchers in psychiatry, nuclear engineering, and radiology. Education: PhD in Electrical Engineering (Rice University, 2004), MS (Rice, 2000), AB in Mathematics (Harvard, 1998). He teaches courses in machine learning (EECS 545), signal processing, and statistical methods. His work emphasizes developing scalable algorithms with theoretical guarantees, particularly in domains like functional neuroimaging, nuclear particle classification, and sepsis prediction. He advises ~1 PhD student annually and has mentored over 20 students. His grants include NSF, NIH, and DOE funding, focusing on topics like domain adaptation, label noise, and medical image registration. His lab develops open-source tools for robust kernel methods, mixture proportion estimation, and partial mixture modeling.
