Randomized algorithms for estimating the trace of an implicit symmetric positive semi-definite matrix.
Haim AvronSivan ToledoPublished in: J. ACM (2011)
Keyphrases
- positive semi definite
- randomized algorithms
- similarity function
- kernel methods
- metric learning
- approximation algorithms
- dot product
- perceptron algorithm
- lower bound
- kernel function
- randomized algorithm
- reproducing kernel hilbert space
- learning algorithm
- kernel matrix
- distance metric
- graph kernels
- worst case
- multi class
- learning tasks
- greedy algorithm
- loss function
- weight vector
- linear combination
- special case