Golub, Gene H. — LLMpedia

Golub, Gene H.
Name	Gene H. Golub
Birth date	March 23, 1932
Birth place	Chișinău, Kingdom of Romania (now Moldova)
Death date	November 16, 2007
Death place	La Jolla, California, United States
Nationality	American
Fields	Numerical analysis, Applied mathematics, Computer science
Workplaces	Stanford University, University of California, Berkeley, University of Michigan
Alma mater	Case Institute of Technology, The Ohio State University
Doctoral advisor	George W. Stewart
Known for	Numerical linear algebra, Golub–Kahan bidiagonalization, singular value decomposition algorithms, matrix computations

Contents

Golub, Gene H. Gene H. Golub was a prominent numerical analyst and computer scientist known for foundational work in numerical linear algebra, algorithm design, and scientific computing. He influenced generations of researchers through seminal textbooks, software libraries, and leadership at institutions and societies spanning Stanford University, University of California, Berkeley, Society for Industrial and Applied Mathematics, and international collaborations with researchers at University of Oxford, ETH Zurich, and Max Planck Society. His work on the singular value decomposition and eigenvalue problems shaped practice at research centers such as Los Alamos National Laboratory, Lawrence Berkeley National Laboratory, and industry labs like IBM Research, Bell Labs, and Microsoft Research.

Early life and education

Born in Chișinău, Golub emigrated to the United States and studied at the Case Institute of Technology, where he earned his bachelor's degree, followed by graduate study at The Ohio State University under advisor George W. Stewart. During his doctoral and postdoctoral years he interacted with scholars from Courant Institute of Mathematical Sciences, California Institute of Technology, and Massachusetts Institute of Technology, and engaged with numerical analysts from institutions including Argonne National Laboratory, National Bureau of Standards, and Princeton University. Early influences included encounters with researchers from Harvard University, Yale University, Columbia University, University of Chicago, and University of Illinois Urbana–Champaign.

Golub held faculty appointments at University of Michigan and later at Stanford University, where he became the Fletcher Jones Professor of Computer Science and Applied Mathematics. He also maintained close ties with University of California, San Diego, University of Cambridge, Imperial College London, and visiting posts at ETH Zurich and University of Kent. Golub served leadership roles in professional organizations including Society for Industrial and Applied Mathematics (SIAM), collaborated with editorial boards for journals such as SIAM Journal on Numerical Analysis, Numerische Mathematik, Journal of the ACM, ACM Transactions on Mathematical Software, and advised students who later worked at Intel, Google, NVIDIA, Oak Ridge National Laboratory, and Sandia National Laboratories. He participated in panels at National Science Foundation, Department of Energy, Office of Naval Research, and international meetings like the International Congress of Mathematicians and the SIAM Annual Meeting.

Golub pioneered algorithms for the computation of the singular value decomposition, eigenvalue problems, and least squares, notably the Golub–Kahan bidiagonalization and stable SVD methods used across physics, engineering, and statistics applications at centers such as CERN, NASA, European Space Agency, and Bell Labs. He coauthored a widely cited textbook with Charles F. Van Loan that influenced curricula at Princeton University, MIT, Stanford University, and University of California, Berkeley. His software contributions underpinned libraries including LINPACK, EISPACK, LAPACK, and influenced implementations in MATLAB, Scilab, R (programming language), NumPy, and SciPy. Golub's work connected to research in numerical optimization at INRIA, data analysis at Los Alamos National Laboratory, and machine learning foundations pursued at Carnegie Mellon University, University of Toronto, and University of Montreal. His students and collaborators include scholars affiliated with Princeton Plasma Physics Laboratory, Brookhaven National Laboratory, Fermi National Accelerator Laboratory, and technology firms such as Apple Inc., Amazon, and Facebook.

Golub received honors from bodies including election to the National Academy of Engineering, the American Academy of Arts and Sciences, and distinctions from Society for Industrial and Applied Mathematics. He was awarded prizes and fellowships connected with National Science Foundation grants, the Guggenheim Fellowship, and recognition from IEEE and ACM. International honors tied him to academies such as the Royal Society (honorary interactions), Deutsche Forschungsgemeinschaft collaborations, and awards from organizations including SIAM and the Mathematical Association of America.

Golub authored and coauthored influential works such as the textbook with Charles F. Van Loan on matrix computations and numerous papers in journals including SIAM Review, Numerische Mathematik, Journal of Computational Physics, and Communications of the ACM. He contributed algorithms and code that fed into EISPACK, LINPACK, LAPACK, and shaped numerical routines used in MATLAB and scientific software at National Institutes of Health laboratories. Key collaborations involved researchers from George Mason University, Rice University, Duke University, University of Washington, and Johns Hopkins University.