Material Detail

Spectral Graph Theory, Linear Solvers and Applications

Spectral Graph Theory, Linear Solvers and Applications

This video was recorded at Machine Learning Summer School (MLSS), Chicago 2009. We discuss the development of combinatorial methods for solving symmetric diagonally dominate linear systems. Over the last fifteen years the computer science community has made substantial progress in fast solvers for SDD systems. For general SDD systems the upper bound is $0(m \log^k n)$ for some constant $k$, where $m$ is the number of non-zero entries, due to Spielman and Teng. Newer methods, combinatorial multigrid, have linear time guarantee for the planar case and work very well in practice. Critical to the use of these new solvers has been the reduction of problems to the solution of SDD systems. We present some of these reductions, including several from image processing.
Rate

Quality

  • User Rating
  • Comments
  • Learning Exercises
  • Bookmark Collections
  • Course ePortfolios
  • Accessibility Info

More about this material

Comments

Log in to participate in the discussions or sign up if you are not already a MERLOT member.