On the Gap Between Strict-Saddles and True Convexity: An Omega(log d) Lower Bound for Eigenvector Approximation.
Max SimchowitzAhmed El AlaouiBenjamin RechtPublished in: CoRR (2017)
Keyphrases
- lower bound
- critical points
- upper bound
- polynomial approximation
- expected error
- branch and bound algorithm
- bayes error rate
- constant factor
- optimal solution
- linear programming relaxation
- branch and bound
- integrality gap
- randomized algorithm
- scale space
- worst case
- approximation error
- closed form
- objective function
- np hard
- partition function
- linear combination
- approximation algorithms
- approximation guarantees
- lower and upper bounds
- lower bounding
- principal component analysis
- spectral clustering
- image analysis
- arbitrarily close
- covariance matrix
- vc dimension
- linear programming
- online algorithms
- knapsack problem
- saddle points
- singular value decomposition