In this talk I will show how to derive the fastest coordinate descent method  and the fastest stochastic gradient descent method , both from the linear-coupling framework . I will relate them to linear system solving, conjugate gradient method, the Chebyshev approximation theory, and raise several open questions at the end. No prior knowledge is required on first-order methods.
Benjamin Elman, Gordon Wu ’58 Professor of Chinese Studies at Princeton University and former Mellon Visiting Professor (1999-2001) in the School of Historical Studies at the Institute, will deliver a public lecture, “Philologists as Rogues? The Life of a Confucian Classic Recovered in Early Modern Japan and Its Transmission Back to Imperial China,” on Friday, November 18, at 4:30 p.m. in West Lecture hall on the Institute campus.