Notes
1. Notes on sensorimotor cortical areas
A rough compilation of material presented during the Sabes' Lab meetings, April-May 2001. Focuses on nomenclature, intracortical connections, and basic physiological properties of various parietal and cortical sensorimotor areas.
2. Notes on "Linear algebraic equations, SVD, and the pseudo-inverse"
These notes were written in 2001 for a graduate level course on applied math taught by Ken Miller. It's based on material taught by Michael Jordan at MIT when I was graduate student.
"Linear Algebraic Equations, SVD, and the Pseudo-Inversee" by Phillip N. Sabes is licensed under a Creative Commons Attribution-Noncommercial-Share Alike 3.0 United States License
Code
1. EM algorithm for linear dynamical systems
A Matlab implementation to identify a linear dynamical system (LDS) of the form
x(t+1) = A*x(t) + B*u(t) + q(t)
y(t) = C*x(t) + D*w(t) + r(t)
cov([q,r])=[Q 0; 0 R]The algorithm is based on the expectation-maximization (EM) algorithm developed by Shumway and Stoffer (1982) and Ghahramani and Hinton (1996). The E-Step consists of Kalman smoothing, the M-Step is solved exactly. This implementation is based on code by Zoubin Ghahramani. Details of this algorithm can be found in the appendix of Cheng and Sabes (2006).
Authors: Philip Sabes and Sen Cheng
License: GNU General Public License
Citation: Cheng S and Sabes PN, Neural Comput., 18(4): 760-793 (2006)
Date Version Notes June 17, 2005 1.0 Initial implementation. Any matrix parameter can be held fixed during estimation.