Reproduce “Marginalised Gaussian Processes with Nested Sampling” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Fergus Simpson, Vidhi Lalchand, Carl Edward Rasmussen
The authors apply nested sampling (NS) to average Gaussian process (GP) forecasts over plausible kernel hyperparameters, rather than selecting one setting with marginal likelihood maximization (ML-II). On 13 time-series tasks with a seven-component spectral mixture kernel, NS achieved lower negative log predictive density than HMC on 11; this advantage did not extend to RBF kernels.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.