Efficient coordinate-wise leading eigenvector computation

Jialei Wang, Weiran Wang, Dan Garber, Nathan Srebro

Research output: Contribution to journalArticlepeer-review


We develop and analyze efficient "coordinate-wise" methods for finding the leading eigenvector, where each step involves only a vector-vector product. We establish global convergence with overall runtime guarantees that are at least as good as Lanczos's method and dominate it for slowly decaying spectrum. Our methods are based on combining a shift-and-invert approach with coordinate-wise algorithms for linear regression.
Original languageAmerican English
JournalarXiv e-prints
StatePublished - 1 Feb 2017


  • coordinate descent
  • eigenvalue problem
  • power method
  • shift-and-invert


Dive into the research topics of 'Efficient coordinate-wise leading eigenvector computation'. Together they form a unique fingerprint.

Cite this