def __init__(self, n_dims, n_ls, n_kt, rng=None):
if rng is None:
self.rng = np.random.RandomState(np.random.randint(0, 10000))
else:
self.rng = rng
# The number of hyperparameters
self.n_dims = n_dims
# The number of lengthscales
self.n_ls = n_ls
# The number of entries of
# kernel matrix for the tasks
self.n_kt = n_kt
# Prior for the Matern52 lengthscales
self.tophat = TophatPrior(-10, 2, rng=self.rng)
# Prior for the covariance amplitude
self.ln_prior = LognormalPrior(mean=0.0, sigma=1.0, rng=self.rng)
# Prior for the noise
self.horseshoe = HorseshoePrior(scale=0.1, rng=self.rng)
self.tophat_task = TophatPrior(-10, 2, rng=self.rng)
def __init__(self, n_dims, n_ls, n_lr, rng=None):
if rng is None:
self.rng = np.random.RandomState(np.random.randint(0, 10000))
else:
self.rng = rng
# The number of hyperparameters
self.n_dims = n_dims
# The number of lengthscales
self.n_ls = n_ls
# The number of params of the bayes linear reg kernel
self.n_lr = n_lr
self.bayes_lin_prior = NormalPrior(sigma=1, mean=0, rng=self.rng)
# Prior for the Matern52 lengthscales
self.tophat = TophatPrior(-10, 2, rng=self.rng)
# Prior for the covariance amplitude
self.ln_prior = LognormalPrior(mean=-2, sigma=1.0, rng=self.rng)
# Prior for the noise
self.horseshoe = HorseshoePrior(scale=0.001, rng=self.rng)
def test(self):
prior = LognormalPrior(sigma=0.1)
# Check sampling
p0 = prior.sample_from_prior(10)
assert len(p0.shape) == 2
assert p0.shape[0] == 10
assert p0.shape[1] == 1
def __init__(self, n_dims):
# The number of hyperparameters
self.n_dims = n_dims
# Prior for the Matern52 lengthscales
self.tophat = TophatPrior(-2, 2)
# Prior for the covariance amplitude
self.ln_prior = LognormalPrior(mean=0.0, sigma=1.0)
# Prior for the noise
self.horseshoe = HorseshoePrior(scale=0.1)
def __init__(self, n_dims, n_ls, n_lr):
# The number of hyperparameters
self.n_dims = n_dims
# The number of lengthscales
self.n_ls = n_ls
# The number of params of the bayes linear reg kernel
self.n_lr = n_lr
# Prior for the Matern52 lengthscales
self.tophat = TophatPrior(-2, 2)
# Prior for the covariance amplitude
self.ln_prior = LognormalPrior(mean=0.0, sigma=1.0)
# Prior for the noise
self.horseshoe = HorseshoePrior(scale=0.1)
def __init__(self, rng=None):
"""
Abstract base class to define the interface for priors
of GP hyperparameter.
Parameters
----------
"""
if rng is None:
self.rng = np.random.RandomState(np.random.randint(0, 10000))
else:
self.rng = rng
# Prior for the alpha
self.ln_prior_alpha = LognormalPrior(sigma=0.1, mean=-10)
# Prior for the sigma^2
#self.ln_prior_beta = LognormalPrior(sigma=0.1, mean=2)
self.horseshoe = HorseshoePrior(scale=0.1)
def __init__(self, n_dims, n_ls, n_kt):
# The number of hyperparameters
self.n_dims = n_dims
# The number of lengthscales
self.n_ls = n_ls
# The number of entries of
# kernel matrix for the tasks
self.n_kt = n_kt
# Prior for the Matern52 lengthscales
self.tophat = TophatPrior(-2, 2)
# Prior for the covariance amplitude
self.ln_prior = LognormalPrior(mean=0.0, sigma=1.0)
# Prior for the noise
self.horseshoe = HorseshoePrior(scale=0.1)
self.tophat_task = TophatPrior(-2, 2)
