gaussian_process.hpp File Reference

Gaussian-Process regression surrogate for Bayesian Optimization – the "gaussian bands" technique named at this campaign's own drafting (Snoek, Larochelle, Adams, "Practical Bayesian Optimization of Machine Learning Algorithms," NeurIPS 2012). More...

#include <cmath>
#include <stdexcept>
#include <vector>
#include "pulsatrix/linear_algebra.hpp"

Include dependency graph for gaussian_process.hpp:

This graph shows which files directly or indirectly include this file:

Go to the source code of this file.

Classes
class	pulsatrix::GaussianProcessRegressor
	A fitted (or queryable-before-fitting-throws) Gaussian Process regressor with a squared-exponential kernel: k(x, x') = sigma_f^2 * exp(-||x - x'||^2 / (2 * length_scale^2)), plus additive observation noise (a small noise_variance is also standard GP practice purely as numerical "jitter" to keep the kernel matrix well-conditioned, independent of whether the underlying objective is actually noisy). More...
struct	pulsatrix::GaussianProcessRegressor::Posterior
	A posterior prediction: mean and variance at one query point. More...

Namespaces
namespace	pulsatrix

Detailed Description

Gaussian-Process regression surrogate for Bayesian Optimization – the "gaussian bands" technique named at this campaign's own drafting (Snoek, Larochelle, Adams, "Practical Bayesian Optimization of Machine Learning Algorithms," NeurIPS 2012).

Note

Zero-mean prior, squared-exponential (RBF) kernel. Posterior mean/variance both reduce to solving a linear system against the training kernel matrix – resolved via SolveLinearSystem (linear_algebra.hpp), reused rather than duplicated; see that file's own header comment for this campaign's Decision Point (c) resolution (Gaussian elimination per query, not a new Cholesky utility, at this campaign's stated small-trial-count scope).