From saunders at stanford.edu Mon Nov 1 14:33:20 2010 From: saunders at stanford.edu (Michael Saunders) Date: Mon Nov 1 14:35:11 2010 Subject: [BANANA] LA/Opt seminar Thursday Nov 4 (Lindsay Erickson) Message-ID: Linear Algebra and Optimization Seminar (CME 510) iCME, Stanford University http://icme.stanford.edu/seminars/seminar.php?seminar_id=2¤t=true 4:15pm Thursday Nov 4, 2010 Y2E2 101 http://campus-map.stanford.edu/index.cfm?ID=04-070 Lindsay Crowl Erickson Senior Member of the Technical Staff Sandia Laboratories, Thermal Fluids/Engineering 8365 lcerick@sandia.gov Lattice Boltzmann-Immersed Boundary Method for Studying Whole Blood Flow Blood is a complex fluid whose intricate dynamics are challenging to capture numerically. This is a computationally expensive problem and one that would benefit greatly from parallelization. The lattice Boltzmann equations describe the behavior of particle distribution functions whose macroscopic behavior mimics that of the Navier-Stokes equations in the correct limit. These equations are highly localized and scale very well in practice. We develop a parallel lattice Boltzmann-Immersed Boundary (LBIB) method to solve the flow dynamics of platelets in whole blood. Platelets play an essential role in blood clotting by adhering to an injury site on a vessel wall and releasing chemicals into the flow. Under arterial flow conditions, these cells have an enhanced concentration near blood vessel walls due to their interaction with red blood cells. We use the LBIB method to analyze the influence of shear rate and hematocrit on lateral platelet motion. Our model shows that the effective diffusion of platelets is significantly lower near the vessel walls compared to the center of the vessel and suggest that this non-uniform diffusion plus an additional radial drift result in a platelet near-wall excess. Forthcoming: Thu 11 Nov David Fong iCME -------------- next part -------------- An HTML attachment was scrubbed... URL: http://csmr.ca.sandia.gov/pipermail/banana/attachments/20101101/1be84b90/attachment.html From mgu at math.berkeley.edu Tue Nov 2 00:39:59 2010 From: mgu at math.berkeley.edu (Ming Gu) Date: Tue Nov 2 00:44:12 2010 Subject: [BANANA] Reminder: LAPACK seminar on Nov. 3, 2010 In-Reply-To: <201010281312.o9SDC0mD019014@panda.math.berkeley.edu> References: <201010281312.o9SDC0mD019014@panda.math.berkeley.edu> Message-ID: Math 290, Section 29, CS 298, Section 6, Fall 2010 (Matrix Computations and Scientific Computing) We meet WEDNESDAYS 11:10 - noon in Room 380 Soda Hall, Berkeley campus. The coordinators are Profs. J. Demmel (demmel@cs.berkeley.edu) and M. Gu (mgu@math.berkeley.edu). The program will be a mixture of research talks and tutorials. The tutorials will provide a partial sequel to Math 221. For the schedule and other details about the seminar, please see math.berkeley.edu/~mgu/LAPACKSeminar.htm Date: Nov. 3, 2010 Speaker: F. Alberto Grunbaum, UC Berkeley Title: Quantum walks and linear algebra Date: Nov. 10, 2010 Speaker: Bin Yu, UC Berkeley Title: Sparse Modeling: a Statistical View From mgu at math.berkeley.edu Wed Nov 3 23:50:36 2010 From: mgu at math.berkeley.edu (Ming Gu) Date: Wed Nov 3 23:54:52 2010 Subject: [BANANA] LAPACK seminar on Nov. 10 Message-ID: <201011040650.oA46oabN017197@panda.math.berkeley.edu> Math 290, Section 29, CS 298, Section 6, Fall 2010 (Matrix Computations and Scientific Computing) We meet WEDNESDAYS 11:10 - noon in Room 380 Soda Hall, Berkeley campus. The coordinators are Profs. J. Demmel (demmel@cs.berkeley.edu) and M. Gu (mgu@math.berkeley.edu). The program will be a mixture of research talks and tutorials. The tutorials will provide a partial sequel to Math 221. For the schedule and other details about the seminar, please see math.berkeley.edu/~mgu/LAPACKSeminar.htm Date: Nov. 10, 2010 Speaker: Bin Yu, UC Berkeley Title: Sparse Modeling: a Statistical View Abstract: Extracting useful information from high-dimensional data is the focus of today's statistical research and practice. After broad success of statistical machine learning on prediction through regularization, interpretability is gaining attention and sparsity has been used as its proxy. With the virtues of both regularization and sparsity, Lasso (L1 penalized L2 minimization) and its extensions have been very popular recently. In this talk, I would like to give an overview on aspects of statistical theory and pratcice of sparse modeling that includes Lasso and its extensions. First, I will explain what useful insights have been learned from model selection consistency analysis of Lasso and an l1 penalized sparse covariance estimation method when p>>n. Second, I will present results on L2-estimation error (when p>>n) and insights learned for a class of M-estimation methods with decomposable penalities. As special cases, our latter results cover Lasso, L1-penalized GLMs, grouped Lasso, and low-rank sparse matrix estimation. (This talk is based on joint works with co-authors Zhao, Meinshausen, Ravikumar, Raskutti, Wainwright, and Neghban.) Date: Nov. 17, 2010 Speaker: Xin Guo, UC Berkeley From saunders at stanford.edu Thu Nov 4 10:21:44 2010 From: saunders at stanford.edu (Michael Saunders) Date: Thu Nov 4 10:23:34 2010 Subject: [BANANA] LA/Opt seminar TODAY (Lindsay Erickson) Message-ID: Reminder: seminar this afternoon Linear Algebra and Optimization Seminar (CME 510) iCME, Stanford University http://icme.stanford.edu/seminars/seminarInfo.php?seminar_id=2 4:15pm Thursday Nov 4, 2010 Y2E2 101 http://campus-map.stanford.edu/index.cfm?ID=04-070 Lindsay Crowl Erickson Senior Member of the Technical Staff Sandia Laboratories, Thermal Fluids/Engineering 8365 lcerick@sandia.gov Lattice Boltzmann-Immersed Boundary Method for Studying Whole Blood Flow Blood is a complex fluid whose intricate dynamics are challenging to capture numerically. This is a computationally expensive problem and one that would benefit greatly from parallelization. The lattice Boltzmann equations describe the behavior of particle distribution functions whose macroscopic behavior mimics that of the Navier-Stokes equations in the correct limit. These equations are highly localized and scale very well in practice. We develop a parallel lattice Boltzmann-Immersed Boundary (LBIB) method to solve the flow dynamics of platelets in whole blood. Platelets play an essential role in blood clotting by adhering to an injury site on a vessel wall and releasing chemicals into the flow. Under arterial flow conditions, these cells have an enhanced concentration near blood vessel walls due to their interaction with red blood cells. We use the LBIB method to analyze the influence of shear rate and hematocrit on lateral platelet motion. Our model shows that the effective diffusion of platelets is significantly lower near the vessel walls compared to the center of the vessel and suggest that this non-uniform diffusion plus an additional radial drift result in a platelet near-wall excess. Next week: Thu 11 Nov David Fong iCME LSMR: An iterative algorithm for sparse least-squares problems -------------- next part -------------- An HTML attachment was scrubbed... URL: http://csmr.ca.sandia.gov/pipermail/banana/attachments/20101104/461a3a23/attachment.html From mgu at math.berkeley.edu Tue Nov 9 00:27:40 2010 From: mgu at math.berkeley.edu (Ming Gu) Date: Tue Nov 9 00:31:52 2010 Subject: [BANANA] Reminder: LAPACK seminar on Nov. 10 In-Reply-To: <14199_1288853690_oA46skgS020195_201011040650.oA46oabN017197@panda.math.berkeley.edu> References: <14199_1288853690_oA46skgS020195_201011040650.oA46oabN017197@panda.math.berkeley.edu> Message-ID: Math 290, Section 29, CS 298, Section 6, Fall 2010 (Matrix Computations and Scientific Computing) We meet WEDNESDAYS 11:10 - noon in Room 380 Soda Hall, Berkeley campus. The coordinators are Profs. J. Demmel (demmel@cs.berkeley.edu) and M. Gu (mgu@math.berkeley.edu). The program will be a mixture of research talks and tutorials. The tutorials will provide a partial sequel to Math 221. For the schedule and other details about the seminar, please see math.berkeley.edu/~mgu/LAPACKSeminar.htm Date: Nov. 10, 2010 Speaker: Bin Yu, UC Berkeley Title: Sparse Modeling: a Statistical View Date: Nov. 17, 2010 Speaker: Xin Guo, UC Berkeley From saunders at stanford.edu Tue Nov 9 11:03:01 2010 From: saunders at stanford.edu (Michael Saunders) Date: Tue Nov 9 11:04:58 2010 Subject: [BANANA] LA/Opt seminar Thursday Nov 11 (David Fong, iCME) Message-ID: Linear Algebra and Optimization Seminar (CME 510) iCME, Stanford University http://icme.stanford.edu/seminars/seminarInfo.php?seminar_id=2 4:15pm Thursday Nov 11, 2010 Y2E2 101 http://campus-map.stanford.edu/index.cfm?ID=04-070 David Fong, iCME PhD student http://www.stanford.edu/~clfong/ LSMR: An iterative algorithm for sparse least-squares problems For 30 years, the standard iterative solver for large rectangular systems Ax ~= b has been LSQR. It is equivalent to applying the conjugate-gradient method to the normal equations A'Ax = A'b. It reduces norm(r) monotonically, where r ~= b - Ax is the residual for the current estimate of x. Our new solver LSMR is equivalent to applying MINRES to the normal equations, so that norm(A'r) decreases monotonically. In practice we observe that norm(r) and a cheap estimate of the backward error are also monotonic. Thus if iterations need to be terminated early, it is safer to use LSMR. LSQR and LSMR are both based on the Golub-Kahan bidiagonalization (a short-term recurrence for generating vectors u and v). Experiments show that if the vectors v are reorthogonalized, the vectors u remain orthogonal, and (almost) vice versa. Matlab, Fortran 90, and python implementations of LSMR are available (www.stanford.edu/group/SOL/software.html ), with local reorthogonalization of v as an option. Plots of various quantities on a range of sparse test problems illustrate the desirable properties of LSMR. Joint work with Michael Saunders. -------------- next part -------------- An HTML attachment was scrubbed... URL: http://csmr.ca.sandia.gov/pipermail/banana/attachments/20101109/79f7f1fa/attachment.html From saunders at stanford.edu Thu Nov 11 10:08:48 2010 From: saunders at stanford.edu (Michael Saunders) Date: Thu Nov 11 10:10:39 2010 Subject: [BANANA] LA/Opt seminar TODAY (David Fong, iCME) Message-ID: Reminder: seminar this afternoon Linear Algebra and Optimization Seminar (CME 510) iCME, Stanford University http://icme.stanford.edu/seminars/seminarInfo.php?seminar_id=2 4:15pm Thursday Nov 11, 2010 Y2E2 101 http://campus-map.stanford.edu/index.cfm?ID=04-070 David Fong, iCME PhD student http://www.stanford.edu/~clfong/ LSMR: An iterative algorithm for sparse least-squares problems For 30 years, the standard iterative solver for large rectangular systems Ax ~= b has been LSQR. It is equivalent to applying the conjugate-gradient method to the normal equations A'Ax = A'b. It reduces norm(r) monotonically, where r ~= b - Ax is the residual for the current estimate of x. Our new solver LSMR is equivalent to applying MINRES to the normal equations, so that norm(A'r) decreases monotonically. In practice we observe that norm(r) and a cheap estimate of the backward error are also monotonic. Thus if iterations need to be terminated early, it is safer to use LSMR. LSQR and LSMR are both based on the Golub-Kahan bidiagonalization (a short-term recurrence for generating vectors u and v). Experiments show that if the vectors v are reorthogonalized, the vectors u remain orthogonal, and (almost) vice versa. Matlab, Fortran 90, and python implementations of LSMR are available (www.stanford.edu/group/SOL/software.html ), with local reorthogonalization of v as an option. Plots of various quantities on a range of sparse test problems illustrate the desirable properties of LSMR. Joint work with Michael Saunders. Next week: Thursday Nov 18 Nick West (iCME) -------------- next part -------------- An HTML attachment was scrubbed... URL: http://csmr.ca.sandia.gov/pipermail/banana/attachments/20101111/16375a50/attachment-0001.html From mgu at math.berkeley.edu Fri Nov 12 22:08:08 2010 From: mgu at math.berkeley.edu (Ming Gu) Date: Fri Nov 12 22:12:24 2010 Subject: [BANANA] LAPACK seminar on Nov. 17 In-Reply-To: <201011040650.oA46oabN017197@panda.math.berkeley.edu> References: <201011040650.oA46oabN017197@panda.math.berkeley.edu> Message-ID: Math 290, Section 29, CS 298, Section 6, Fall 2010 (Matrix Computations and Scientific Computing) We meet WEDNESDAYS 11:10 - noon in Room 380 Soda Hall, Berkeley campus. The coordinators are Profs. J. Demmel (demmel@cs.berkeley.edu) and M. Gu (mgu@math.berkeley.edu). The program will be a mixture of research talks and tutorials. The tutorials will provide a partial sequel to Math 221. For the schedule and other details about the seminar, please see math.berkeley.edu/~mgu/LAPACKSeminar.htm Date: Nov. 17, 2010 Speaker: Xin Guo, UC Berkeley Date: Nov. 24, 2010 Speaker: Lek-Heng Lim, University of Chicago From mgu at math.berkeley.edu Mon Nov 15 20:35:14 2010 From: mgu at math.berkeley.edu (Ming Gu) Date: Mon Nov 15 20:39:28 2010 Subject: [BANANA] LAPACK seminar on Nov. 17, 2010 Message-ID: <201011160435.oAG4ZEMM001967@panda.math.berkeley.edu> Math 290, Section 29, CS 298, Section 6, Fall 2010 (Matrix Computations and Scientific Computing) We meet WEDNESDAYS 11:10 - noon in Room 380 Soda Hall, Berkeley campus. The coordinators are Profs. J. Demmel (demmel@cs.berkeley.edu) and M. Gu (mgu@math.berkeley.edu). The program will be a mixture of research talks and tutorials. The tutorials will provide a partial sequel to Math 221. For the schedule and other details about the seminar, please see math.berkeley.edu/~mgu/LAPACKSeminar.htm Date: Nov. 17 Speaker: Xin Guo, UC Berkeley Title: Algebraic methods for modeling default dependence Abstract: The past/current financial crisis highlights the major issue of correlated default. In this talk we will first give an overview of standard approaches in modeling correlated default. We will then present a graphic model for default dependence. We will show how this model is connected to well-known results in algebraic statistics and discuss the computational complexity of this type of models. Based on joint work with O. I. Filiz, J. Morton and B. Sturmfels. Date: Nov. 24, 2010 Speaker: Lek-Heng Lim, University of Chicago From saunders at stanford.edu Tue Nov 16 16:13:14 2010 From: saunders at stanford.edu (Michael Saunders) Date: Tue Nov 16 16:15:10 2010 Subject: [BANANA] LA/Opt seminar Thursday (Nick West) Message-ID: Linear Algebra and Optimization Seminar (CME 510) iCME, Stanford University http://icme.stanford.edu/seminars/seminarInfo.php?seminar_id=2 4:15pm Thursday Nov 18, 2010 Y2E2 101 http://campus-map.stanford.edu/index.cfm?ID=04-070 Nick West, iCME PhD candidate nickwest@stanford.edu http://nickwest.stanford.edu Solution of the Helmholtz Equation in Spherical Coordinates We are interested in studying the solution of the variable coefficient Helmholtz equation in spherical coordinates that arises in ocean modeling. The use of spherical coordinates (as dictated by the geometry of the earth) leads to large anisotropy in the operator near the poles, causing slow convergence in these regions. Previous work [Brown, 2004] addressed this problem through the use of preconditioners; we approach the problem differently. We treat the Helmholtz equation in spherical coordinates as a variable coefficient elliptic equation in Cartesian coordinates and, following Concus and Golub [1973], we introduce a change of variables that reduces the problem to solving a much simpler set of equations. These modified equations suggest a natural splitting for an iterative method, and cyclic reduction is used to greatly reduce the size of the system to be solved at each iteration and to parallelize the solve. We find that this method converges extremely quickly in the transformed variables, but the residual is somewhat larger in the original variables (though for the ocean modeling application, the solution is accurate enough). This work was done jointly with Felix Kwok, Gene Golub, and Nancy Nichols in 2006. -------------- next part -------------- An HTML attachment was scrubbed... URL: http://csmr.ca.sandia.gov/pipermail/banana/attachments/20101116/0340d2ce/attachment.html From saunders at stanford.edu Thu Nov 18 09:49:41 2010 From: saunders at stanford.edu (Michael Saunders) Date: Thu Nov 18 09:51:35 2010 Subject: [BANANA] LA/Opt seminar TODAY (Nick West) Message-ID: Reminder: Seminar this afternoon Linear Algebra and Optimization Seminar (CME 510) iCME, Stanford University http://icme.stanford.edu/seminars/seminarInfo.php?seminar_id=2 4:15pm Thursday Nov 18, 2010 Y2E2 101 http://campus-map.stanford.edu/index.cfm?ID=04-070 Nick West, iCME PhD candidate nickwest@stanford.edu http://nickwest.stanford.edu Solution of the Helmholtz Equation in Spherical Coordinates We are interested in studying the solution of the variable coefficient Helmholtz equation in spherical coordinates that arises in ocean modeling. The use of spherical coordinates (as dictated by the geometry of the earth) leads to large anisotropy in the operator near the poles, causing slow convergence in these regions. Previous work [Brown, 2004] addressed this problem through the use of preconditioners; we approach the problem differently. We treat the Helmholtz equation in spherical coordinates as a variable coefficient elliptic equation in Cartesian coordinates and, following Concus and Golub [1973], we introduce a change of variables that reduces the problem to solving a much simpler set of equations. These modified equations suggest a natural splitting for an iterative method, and cyclic reduction is used to greatly reduce the size of the system to be solved at each iteration and to parallelize the solve. We find that this method converges extremely quickly in the transformed variables, but the residual is somewhat larger in the original variables (though for the ocean modeling application, the solution is accurate enough). This work was done jointly with Felix Kwok, Gene Golub, and Nancy Nichols in 2006. This is (probably) the last LA/Opt seminar of Fall quarter. -------------- next part -------------- An HTML attachment was scrubbed... URL: http://csmr.ca.sandia.gov/pipermail/banana/attachments/20101118/c449d746/attachment.html From mgu at math.berkeley.edu Sun Nov 21 12:36:44 2010 From: mgu at math.berkeley.edu (Ming Gu) Date: Sun Nov 21 12:41:04 2010 Subject: [BANANA] LAPACK seminar on Nov. 24, 2010 Message-ID: <201011212036.oALKaio2017639@panda.math.berkeley.edu> Math 290, Section 29, CS 298, Section 6, Fall 2010 (Matrix Computations and Scientific Computing) We meet WEDNESDAYS 11:10 - noon in Room 380 Soda Hall, Berkeley campus. The coordinators are Profs. J. Demmel (demmel@cs.berkeley.edu) and M. Gu (mgu@math.berkeley.edu). The program will be a mixture of research talks and tutorials. The tutorials will provide a partial sequel to Math 221. For the schedule and other details about the seminar, please see math.berkeley.edu/~mgu/LAPACKSeminar.htm Date: Nov. 24, 2010 Speaker: Lek-Heng Lim, University of Chicago Title: Cholesky Decomposable Tensors Abstract: A real quadratic form f(x) = x'Ax may be expressed as a sum of squares of linear forms f(x) = \sum_k (b_k'x)^2 iff the associated symmetric matrix A is positive semidefinite. We will examine a generalization of this notion: degree-p forms that are expressible as a sum of p-powers of linear forms: f(x) = \sum_k (b_k'x)^p (p even). In this case it is not sufficient that the symmetric tensor associated with f be positive semidefinite although that is a necessary condition. We call such tensors Cholesky decomposable tensors and show that they are surprisingly well-behaved. The sum-of-powers decomposition is then a higher-order analogue of the spectral decomposition of a positive semidefinite matrix. It arises naturally in diffusion MRI and gives the most accurate method for extracting nerve fibers crossing. This is joint work with T. Schultz. Date: Dec. 1, 2010 Speaker: Cinna Wu, UC Berkeley From mgu at math.berkeley.edu Sun Nov 21 12:42:39 2010 From: mgu at math.berkeley.edu (Ming Gu) Date: Sun Nov 21 12:46:52 2010 Subject: [BANANA] LAPACK seminar on Nov. 24, 2010 Message-ID: <201011212042.oALKgdiK017706@panda.math.berkeley.edu> Math 290, Section 29, CS 298, Section 6, Fall 2010 (Matrix Computations and Scientific Computing) We meet WEDNESDAYS 11:10 - noon in Room 380 Soda Hall, Berkeley campus. The coordinators are Profs. J. Demmel (demmel@cs.berkeley.edu) and M. Gu (mgu@math.berkeley.edu). The program will be a mixture of research talks and tutorials. The tutorials will provide a partial sequel to Math 221. For the schedule and other details about the seminar, please see math.berkeley.edu/~mgu/LAPACKSeminar.htm Date: Nov. 24, 2010 Speaker: Lek-Heng Lim, University of Chicago Title: Cholesky Decomposable Tensors Abstract: A real quadratic form f(x) = x'Ax may be expressed as a sum of squares of linear forms f(x) = \sum_k (b_k'x)^2 iff the associated symmetric matrix A is positive semidefinite. We will examine a generalization of this notion: degree-p forms that are expressible as a sum of p-powers of linear forms: f(x) = \sum_k (b_k'x)^p (p even). In this case it is not sufficient that the symmetric tensor associated with f be positive semidefinite although that is a necessary condition. We call such tensors Cholesky decomposable tensors and show that they are surprisingly well-behaved. The sum-of-powers decomposition is then a higher-order analogue of the spectral decomposition of a positive semidefinite matrix. It arises naturally in diffusion MRI and gives the most accurate method for extracting nerve fibers crossing. This is joint work with T. Schultz. Date: Dec. 1, 2010 Speaker: Cinna Wu From mgu at math.berkeley.edu Tue Nov 23 00:38:34 2010 From: mgu at math.berkeley.edu (Ming Gu) Date: Tue Nov 23 00:42:48 2010 Subject: [BANANA] Reminder: LAPACK seminar on Nov. 24, 2010 In-Reply-To: <32301_1290372061_oALKewDI014964_201011212036.oALKaio2017639@panda.math.berkeley.edu> References: <32301_1290372061_oALKewDI014964_201011212036.oALKaio2017639@panda.math.berkeley.edu> Message-ID: Math 290, Section 29, CS 298, Section 6, Fall 2010 (Matrix Computations and Scientific Computing) We meet WEDNESDAYS 11:10 - noon in Room 380 Soda Hall, Berkeley campus. The coordinators are Profs. J. Demmel (demmel@cs.berkeley.edu) and M. Gu (mgu@math.berkeley.edu). The program will be a mixture of research talks and tutorials. The tutorials will provide a partial sequel to Math 221. For the schedule and other details about the seminar, please see math.berkeley.edu/~mgu/LAPACKSeminar.htm Date: Nov. 24, 2010 Speaker: Lek-Heng Lim, University of Chicago Title: Cholesky Decomposable Tensors Date: Dec. 1, 2010 Speaker: Cinna Wu, UC Berkeley From mgu at math.berkeley.edu Mon Nov 29 00:04:25 2010 From: mgu at math.berkeley.edu (Ming Gu) Date: Mon Nov 29 00:08:40 2010 Subject: [BANANA] LAPACK seminar on Dec. 1, 2010 Message-ID: <201011290804.oAT84PcM002093@panda.math.berkeley.edu> Math 290, Section 29, CS 298, Section 6, Fall 2010 (Matrix Computations and Scientific Computing) We meet WEDNESDAYS 11:10 - noon in Room 380 Soda Hall, Berkeley campus. The coordinators are Profs. J. Demmel (demmel@cs.berkeley.edu) and M. Gu (mgu@math.berkeley.edu). The program will be a mixture of research talks and tutorials. The tutorials will provide a partial sequel to Math 221. For the schedule and other details about the seminar, please see math.berkeley.edu/~mgu/LAPACKSeminar.htm Date: Dec. 1, 2010 Speaker: Cinna Wu, UC Berkeley Title: An Introduction to Structured Low Rank Matrix Approximation Date: Dec. 8, 2010 Speaker: Yuji Nakatsukasa, UC Davis From mgu at math.berkeley.edu Tue Nov 30 14:27:52 2010 From: mgu at math.berkeley.edu (Ming Gu) Date: Tue Nov 30 14:32:04 2010 Subject: [BANANA] Reminder: LAPACK seminar on Dec. 1, 2010 In-Reply-To: <201011290804.oAT84PcM002093@panda.math.berkeley.edu> References: <201011290804.oAT84PcM002093@panda.math.berkeley.edu> Message-ID: Math 290, Section 29, CS 298, Section 6, Fall 2010 (Matrix Computations and Scientific Computing) We meet WEDNESDAYS 11:10 - noon in Room 380 Soda Hall, Berkeley campus. The coordinators are Profs. J. Demmel (demmel@cs.berkeley.edu) and M. Gu (mgu@math.berkeley.edu). The program will be a mixture of research talks and tutorials. The tutorials will provide a partial sequel to Math 221. For the schedule and other details about the seminar, please see math.berkeley.edu/~mgu/LAPACKSeminar.htm Date: Dec. 1, 2010 Speaker: Cinna Wu, UC Berkeley Title: An Introduction to Structured Low Rank Matrix Approximation Date: Dec. 8, 2010 Speaker: Yuji Nakatsukasa, UC Davis