Name
Abstract | ||
|---|---|---|
We propose Kernel Hamiltonian Monte Carlo (KMC), a gradient-free adaptive MCMC algorithm based on Hamiltonian Monte Carlo (HMC). On target densities where classical HMC is not an option due to intractable gradients, KMC adaptively learns the target's gradient structure by fitting an exponential family model in a Reproducing Kernel Hilbert Space. Computational costs are reduced by two novel efficient approximations to this gradient. While being asymptotically exact, KMC mimics HMC in terms of sampling efficiency, and offers substantial mixing improvements over state-of-the-art gradient free samplers. We support our claims with experimental studies on both toy and real-world applications, including Approximate Bayesian Computation and exact-approximate MCMC. |
Year | Venue | Field |
|---|---|---|
2015 | ADVANCES IN NEURAL INFORMATION PROCESSING SYSTEMS 28 (NIPS 2015) | Kernel (linear algebra),Applied mathematics,Mathematical optimization,Approximate Bayesian computation,Markov chain Monte Carlo,Exponential family,Hybrid Monte Carlo,Theoretical computer science,Sampling (statistics),Mathematics,Reproducing kernel Hilbert space |
DocType | Volume | ISSN |
Conference | 28 | 1049-5258 |
Citations | PageRank | References |
3 | 0.43 | 12 |
Authors | ||
5 | ||
Name | Order | Citations | PageRank |
|---|---|---|---|
| Heiko Strathmann | 1 | 82 | 5.84 |
| Dino Sejdinovic | 2 | 443 | 37.96 |
| Samuel Livingstone | 3 | 8 | 1.14 |
| Zoltán Szabó | 4 | 3 | 0.43 |
| Arthur Gretton | 5 | 3638 | 226.18 |