Super-Samples from Kernel Herding
arXiv:1203.3472
Abstract
We extend the herding algorithm to continuous spaces by using the kernel trick. The resulting "kernel herding" algorithm is an infinite memory deterministic process that learns to approximate a PDF with a collection of samples. We show that kernel herding decreases the error of expectations of functions in the Hilbert space at a rate O(1/T) which is much faster than the usual O(1/pT) for iid random samples. We illustrate kernel herding by approximating Bayesian predictive distributions.
Appears in Proceedings of the Twenty-Sixth Conference on Uncertainty in Artificial Intelligence (UAI2010)
Cited by in corpus (6)
- Good Subnetworks Provably Exist: Pruning via Greedy Forward Selection
- Active Testing: Sample-Efficient Model Evaluation
- A Kernel Mean Embedding Approach to Reducing Conservativeness in Stochastic Programming and Control
- Herding Generalizes Diverse M -Best Solutions
- Tuned Compositional Feature Replays for Efficient Stream Learning
- Density Sketches for Sampling and Estimation