#!/usr/bin/env python3
# Copyright (c) Meta Platforms, Inc. and affiliates.
#
# This source code is licensed under the MIT license found in the
# LICENSE file in the root directory of this source tree.
r"""
Batch acquisition functions using the reparameterization trick in combination
with (quasi) Monte-Carlo sampling. See [Rezende2014reparam]_, [Wilson2017reparam]_ and
[Balandat2020botorch]_.
References
.. [Rezende2014reparam]
D. J. Rezende, S. Mohamed, and D. Wierstra. Stochastic backpropagation and
approximate inference in deep generative models. ICML 2014.
.. [Wilson2017reparam]
J. T. Wilson, R. Moriconi, F. Hutter, and M. P. Deisenroth.
The reparameterization trick for acquisition functions. ArXiv 2017.
"""
from __future__ import annotations
import math
from abc import ABC, abstractmethod
from collections.abc import Callable
from copy import deepcopy
from functools import partial
from typing import Protocol
import torch
from botorch.acquisition.acquisition import AcquisitionFunction, MCSamplerMixin
from botorch.acquisition.cached_cholesky import CachedCholeskyMCSamplerMixin
from botorch.acquisition.objective import (
ConstrainedMCObjective,
IdentityMCObjective,
MCAcquisitionObjective,
PosteriorTransform,
)
from botorch.acquisition.utils import (
compute_best_feasible_objective,
prune_inferior_points,
repeat_to_match_aug_dim,
)
from botorch.exceptions.errors import UnsupportedError
from botorch.exceptions.warnings import legacy_ei_numerics_warning
from botorch.models.model import Model
from botorch.sampling.base import MCSampler
from botorch.utils.objective import compute_smoothed_feasibility_indicator
from botorch.utils.transforms import (
concatenate_pending_points,
match_batch_shape,
t_batch_mode_transform,
)
from torch import Tensor
[docs]
class MCAcquisitionFunction(AcquisitionFunction, MCSamplerMixin, ABC):
r"""
Abstract base class for Monte-Carlo based batch acquisition functions.
"""
def __init__(
self,
model: Model,
sampler: MCSampler | None = None,
objective: MCAcquisitionObjective | None = None,
posterior_transform: PosteriorTransform | None = None,
X_pending: Tensor | None = None,
) -> None:
r"""
Args:
model: A fitted model.
sampler: The sampler used to draw base samples. If not given,
a sampler is generated on the fly within the
`get_posterior_samples` method using
`botorch.sampling.get_sampler`.
NOTE: For posteriors that do not support base samples,
a sampler compatible with intended use case must be provided.
See `ForkedRNGSampler` and `StochasticSampler` as examples.
objective: The MCAcquisitionObjective under which the samples are
evaluated. Defaults to `IdentityMCObjective()`.
posterior_transform: A PosteriorTransform (optional).
X_pending: A `batch_shape, m x d`-dim Tensor of `m` design points
that have points that have been submitted for function evaluation
but have not yet been evaluated.
"""
super().__init__(model=model)
MCSamplerMixin.__init__(self, sampler=sampler)
if objective is None and model.num_outputs != 1:
if posterior_transform is None:
raise UnsupportedError(
"Must specify an objective or a posterior transform when using "
"a multi-output model."
)
elif not posterior_transform.scalarize:
raise UnsupportedError(
"If using a multi-output model without an objective, "
"posterior_transform must scalarize the output."
)
if objective is None:
objective = IdentityMCObjective()
self.posterior_transform = posterior_transform
self.objective: MCAcquisitionObjective = objective
self.set_X_pending(X_pending)
def _get_samples_and_objectives(self, X: Tensor) -> tuple[Tensor, Tensor]:
"""Computes posterior samples and objective values at input X.
Args:
X: A `batch_shape x q x d`-dim Tensor of model inputs.
Returns:
A two-tuple `(samples, obj)`, where `samples` is a tensor of posterior
samples with shape `sample_shape x batch_shape x q x m`, and `obj` is a
tensor of MC objective values with shape `sample_shape x batch_shape x q`.
"""
posterior = self.model.posterior(
X=X, posterior_transform=self.posterior_transform
)
samples = self.get_posterior_samples(posterior)
obj = self.objective(samples=samples, X=X)
return samples, obj
[docs]
@abstractmethod
def forward(self, X: Tensor) -> Tensor:
r"""Takes in a `batch_shape x q x d` X Tensor of t-batches with `q` `d`-dim
design points each, and returns a Tensor with shape `batch_shape'`, where
`batch_shape'` is the broadcasted batch shape of model and input `X`. Should
utilize the result of `set_X_pending` as needed to account for pending function
evaluations.
"""
pass # pragma: no cover
[docs]
class SampleReductionProtocol(Protocol):
"""For static type check of SampleReducingMCAcquisitionFunction's mc_reduction."""
@staticmethod
def __call__(X: Tensor, *, dim: torch.Size) -> Tensor:
pass # pragma: no cover
[docs]
class SampleReducingMCAcquisitionFunction(MCAcquisitionFunction):
r"""MC-based batch acquisition function that reduces across samples and implements
a general treatment of outcome constraints.
This class's `forward` computes the - possibly constrained - acquisition value by
(1) computing the unconstrained utility for each MC sample using `_sample_forward`,
(2) weighing the utility values by the constraint indicator per MC sample, and
(3) reducing (e.g. averaging) the weighted utility values over the MC dimension.
NOTE: Do *NOT* override the `forward` method, unless you have thought about it well.
`forward` is implemented generically to incorporate constraints in a principled way,
and takes care of reducing over the Monte Carlo and batch dimensions via the
`sample_reduction` and `q_reduction` arguments, which default to `torch.mean` and
`torch.max`, respectively.
In order to implement a custom SampleReducingMCAcquisitionFunction, we only need to
implement the `_sample_forward(obj: Tensor) -> Tensor` method, which maps objective
samples to acquisition utility values without reducing the Monte Carlo and batch
(i.e. q) dimensions (see details in the docstring of `_sample_forward`).
A note on design choices:
The primary purpose of `SampleReducingMCAcquisitionFunction`is to support outcome
constraints. On the surface, designing a wrapper `ConstrainedMCAcquisitionFunction`
could be an elegant solution to this end, but it would still require the acquisition
functions to implement a `_sample_forward` method to weigh acquisition utilities at
the sample level. Further, `qNoisyExpectedImprovement` is a special case that is
hard to encompass in this pattern, since it requires the computation of the best
*feasible* objective, which requires access to the constraint functions. However,
if the constraints are stored in a wrapper class, they will be inaccessible to the
forward pass. These problems are circumvented by the design of this class.
"""
_log: bool = False # whether the acquisition utilities are in log-space
def __init__(
self,
model: Model,
sampler: MCSampler | None = None,
objective: MCAcquisitionObjective | None = None,
posterior_transform: PosteriorTransform | None = None,
X_pending: Tensor | None = None,
sample_reduction: SampleReductionProtocol = torch.mean,
q_reduction: SampleReductionProtocol = torch.amax,
constraints: list[Callable[[Tensor], Tensor]] | None = None,
eta: Tensor | float = 1e-3,
fat: bool = False,
):
r"""Constructor of SampleReducingMCAcquisitionFunction.
Args:
model: A fitted model.
sampler: The sampler used to draw base samples. If not given, a
sampler is generated on the fly within the
`get_posterior_samples` method using
`botorch.sampling.get_sampler`.
NOTE: For posteriors that do not support base samples,
a sampler compatible with intended use case must be provided.
See `ForkedRNGSampler` and `StochasticSampler` as examples.
objective: The MCAcquisitionObjective under which the samples are
evaluated. Defaults to `IdentityMCObjective()`.
NOTE: `ConstrainedMCObjective` for outcome constraints is deprecated in
favor of passing the `constraints` directly to this constructor.
posterior_transform: A `PosteriorTransform` (optional).
X_pending: A `batch_shape, m x d`-dim Tensor of `m` design points
that have points that have been submitted for function evaluation
but have not yet been evaluated.
sample_reduction: A callable that takes in a `sample_shape x batch_shape`
Tensor of acquisition utility values, a keyword-argument `dim` that
specifies the sample dimensions to reduce over, and returns a
`batch_shape`-dim Tensor of acquisition values.
q_reduction: A callable that takes in a `sample_shape x batch_shape x q`
Tensor of acquisition utility values, a keyword-argument `dim` that
specifies the q dimension to reduce over (i.e. -1), and returns a
`sample_shape x batch_shape`-dim Tensor of acquisition values.
constraints: A list of constraint callables which map a Tensor of posterior
samples of dimension `sample_shape x batch-shape x q x m`-dim to a
`sample_shape x batch-shape x q`-dim Tensor. The associated constraints
are considered satisfied if the output is less than zero.
NOTE: Constraint-weighting is only compatible with non-negative
acquistion utilities, e.g. all improvement-based acquisition functions.
eta: Temperature parameter(s) governing the smoothness of the sigmoid
approximation to the constraint indicators. For more details, on this
parameter, see the docs of `compute_smoothed_feasibility_indicator`.
fat: Wether to apply a fat-tailed smooth approximation to the feasibility
indicator or the canonical sigmoid approximation.
"""
if constraints is not None and isinstance(objective, ConstrainedMCObjective):
raise ValueError(
"ConstrainedMCObjective as well as constraints passed to constructor."
"Choose one or the other, preferably the latter."
)
# TODO: deprecate ConstrainedMCObjective
super().__init__(
model=model,
sampler=sampler,
objective=objective,
posterior_transform=posterior_transform,
X_pending=X_pending,
)
# Shall the need arise, sample_dim could be exposed in the constructor.
sample_dim = tuple(range(len(self.sample_shape)))
self._sample_reduction = partial(sample_reduction, dim=sample_dim)
self._q_reduction = partial(q_reduction, dim=-1)
self._constraints = constraints
self._eta = eta
self._fat = fat
[docs]
@concatenate_pending_points
@t_batch_mode_transform()
def forward(self, X: Tensor) -> Tensor:
r"""Computes the acquisition value associated with the input `X`. Weighs the
acquisition utility values by smoothed constraint indicators if `constraints`
was passed to the constructor of the class. Applies `self.sample_reduction` and
`self.q_reduction` to reduce over the Monte Carlo and batch (q) dimensions.
NOTE: Do *NOT* override the `forward` method for a custom acquisition function.
Instead, implement the `_sample_forward` method. See the docstring of this class
for details.
Args:
X: A `batch_shape x q x d` Tensor of t-batches with `q` `d`-dim
design points each.
Returns:
A Tensor with shape `batch_shape'`, where `batch_shape'` is the broadcasted
batch shape of model and input `X`.
"""
non_reduced_acqval = self._non_reduced_forward(X=X)
return self._sample_reduction(self._q_reduction(non_reduced_acqval))
def _non_reduced_forward(self, X: Tensor) -> Tensor:
"""Compute the constrained acquisition values at the MC-sample, q level.
Args:
X: A `batch_shape x q x d` Tensor of t-batches with `q` `d`-dim
design points each.
Returns:
A Tensor with shape `sample_sample x batch_shape x q`.
"""
samples, obj = self._get_samples_and_objectives(X)
samples = repeat_to_match_aug_dim(target_tensor=samples, reference_tensor=obj)
acqval = self._sample_forward(obj) # `sample_sample x batch_shape x q`
return self._apply_constraints(acqval=acqval, samples=samples)
@abstractmethod
def _sample_forward(self, obj: Tensor) -> Tensor:
"""Evaluates the acquisition utility per MC sample based on objective value obj.
Should utilize the result of `set_X_pending` as needed to account for pending
function evaluations.
Args:
obj: A `sample_shape x batch_shape x q`-dim Tensor of MC objective values.
Returns:
A `sample_shape x batch_shape x q`-dim Tensor of acquisition utility values.
"""
pass # pragma: no cover
def _apply_constraints(self, acqval: Tensor, samples: Tensor) -> Tensor:
"""Multiplies the acquisition utility by constraint indicators.
Args:
acqval: `sample_shape x batch_shape x q`-dim acquisition utility values.
samples: `sample_shape x batch_shape x q x m`-dim posterior samples.
Returns:
A `sample_shape x batch_shape x q`-dim Tensor of acquisition utility values
multiplied by a smoothed constraint indicator per sample.
"""
if self._constraints is not None:
if not self._log and (acqval < 0).any():
raise ValueError(
"Constraint-weighting requires unconstrained "
"acquisition values to be non-negative."
)
ind = compute_smoothed_feasibility_indicator(
constraints=self._constraints,
samples=samples,
eta=self._eta,
log=self._log,
fat=self._fat,
)
acqval = acqval.add(ind) if self._log else acqval.mul(ind)
return acqval
[docs]
class qExpectedImprovement(SampleReducingMCAcquisitionFunction):
r"""MC-based batch Expected Improvement.
This computes qEI by
(1) sampling the joint posterior over q points
(2) evaluating the improvement over the current best for each sample
(3) maximizing over q
(4) averaging over the samples
`qEI(X) = E(max(max Y - best_f, 0)), Y ~ f(X), where X = (x_1,...,x_q)`
Example:
>>> model = SingleTaskGP(train_X, train_Y)
>>> best_f = train_Y.max()[0]
>>> sampler = SobolQMCNormalSampler(1024)
>>> qEI = qExpectedImprovement(model, best_f, sampler)
>>> qei = qEI(test_X)
NOTE: It is strongly recommended to use qLogExpectedImprovement instead
of regular qEI, as it can lead to substantially improved BO performance through
improved numerics. See https://arxiv.org/abs/2310.20708 for details.
"""
def __init__(
self,
model: Model,
best_f: float | Tensor,
sampler: MCSampler | None = None,
objective: MCAcquisitionObjective | None = None,
posterior_transform: PosteriorTransform | None = None,
X_pending: Tensor | None = None,
constraints: list[Callable[[Tensor], Tensor]] | None = None,
eta: Tensor | float = 1e-3,
) -> None:
r"""q-Expected Improvement.
Args:
model: A fitted model.
best_f: The best objective value observed so far (assumed noiseless). Can be
a scalar, or a `batch_shape`-dim tensor. In case of a batched model, the
tensor can specify different values for each element of the batch.
sampler: The sampler used to draw base samples. See `MCAcquisitionFunction`
more details.
objective: The MCAcquisitionObjective under which the samples are evaluated.
Defaults to `IdentityMCObjective()`.
NOTE: `ConstrainedMCObjective` for outcome constraints is deprecated in
favor of passing the `constraints` directly to this constructor.
posterior_transform: A PosteriorTransform (optional).
X_pending: A `m x d`-dim Tensor of `m` design points that have been
submitted for function evaluation but have not yet been evaluated.
Concatenated into X upon forward call. Copied and set to have no
gradient.
constraints: A list of constraint callables which map a Tensor of posterior
samples of dimension `sample_shape x batch-shape x q x m`-dim to a
`sample_shape x batch-shape x q`-dim Tensor. The associated constraints
are considered satisfied if the output is less than zero.
eta: Temperature parameter(s) governing the smoothness of the sigmoid
approximation to the constraint indicators. For more details, on this
parameter, see the docs of `compute_smoothed_feasibility_indicator`.
"""
legacy_ei_numerics_warning(legacy_name=type(self).__name__)
super().__init__(
model=model,
sampler=sampler,
objective=objective,
posterior_transform=posterior_transform,
X_pending=X_pending,
constraints=constraints,
eta=eta,
)
self.register_buffer("best_f", torch.as_tensor(best_f, dtype=float))
def _sample_forward(self, obj: Tensor) -> Tensor:
r"""Evaluate qExpectedImprovement per sample on the candidate set `X`.
Args:
obj: A `sample_shape x batch_shape x q`-dim Tensor of MC objective values.
Returns:
A `sample_shape x batch_shape x q`-dim Tensor of improvement utility values.
"""
return (obj - self.best_f.unsqueeze(-1).to(obj)).clamp_min(0)
[docs]
class qNoisyExpectedImprovement(
SampleReducingMCAcquisitionFunction, CachedCholeskyMCSamplerMixin
):
r"""MC-based batch Noisy Expected Improvement.
This function does not assume a `best_f` is known (which would require
noiseless observations). Instead, it uses samples from the joint posterior
over the `q` test points and previously observed points. The improvement
over previously observed points is computed for each sample and averaged.
`qNEI(X) = E(max(max Y - max Y_baseline, 0))`, where
`(Y, Y_baseline) ~ f((X, X_baseline)), X = (x_1,...,x_q)`
Example:
>>> model = SingleTaskGP(train_X, train_Y)
>>> sampler = SobolQMCNormalSampler(1024)
>>> qNEI = qNoisyExpectedImprovement(model, train_X, sampler)
>>> qnei = qNEI(test_X)
NOTE: It is strongly recommended to use qLogNoisyExpectedImprovement instead
of regular qNEI, as it can lead to substantially improved BO performance through
improved numerics. See https://arxiv.org/abs/2310.20708 for details.
"""
def __init__(
self,
model: Model,
X_baseline: Tensor,
sampler: MCSampler | None = None,
objective: MCAcquisitionObjective | None = None,
posterior_transform: PosteriorTransform | None = None,
X_pending: Tensor | None = None,
prune_baseline: bool = True,
cache_root: bool = True,
constraints: list[Callable[[Tensor], Tensor]] | None = None,
eta: Tensor | float = 1e-3,
marginalize_dim: int | None = None,
) -> None:
r"""q-Noisy Expected Improvement.
Args:
model: A fitted model.
X_baseline: A `batch_shape x r x d`-dim Tensor of `r` design points
that have already been observed. These points are considered as
the potential best design point.
sampler: The sampler used to draw base samples. See `MCAcquisitionFunction`
more details.
objective: The MCAcquisitionObjective under which the samples are
evaluated. Defaults to `IdentityMCObjective()`.
NOTE: `ConstrainedMCObjective` for outcome constraints is deprecated in
favor of passing the `constraints` directly to this constructor.
posterior_transform: A PosteriorTransform (optional).
X_pending: A `batch_shape x m x d`-dim Tensor of `m` design points
that have points that have been submitted for function evaluation
but have not yet been evaluated. Concatenated into `X` upon
forward call. Copied and set to have no gradient.
prune_baseline: If True, remove points in `X_baseline` that are
highly unlikely to be the best point. This can significantly
improve performance and is generally recommended. In order to
customize pruning parameters, instead manually call
`botorch.acquisition.utils.prune_inferior_points` on `X_baseline`
before instantiating the acquisition function.
cache_root: A boolean indicating whether to cache the root
decomposition over `X_baseline` and use low-rank updates.
constraints: A list of constraint callables which map a Tensor of posterior
samples of dimension `sample_shape x batch-shape x q x m`-dim to a
`sample_shape x batch-shape x q`-dim Tensor. The associated constraints
are considered satisfied if the output is less than zero.
eta: Temperature parameter(s) governing the smoothness of the sigmoid
approximation to the constraint indicators. For more details, on this
parameter, see the docs of `compute_smoothed_feasibility_indicator`.
marginalize_dim: The dimension to marginalize over.
TODO: similar to qNEHVI, when we are using sequential greedy candidate
selection, we could incorporate pending points X_baseline and compute
the incremental qNEI from the new point. This would greatly increase
efficiency for large batches.
"""
legacy_ei_numerics_warning(legacy_name=type(self).__name__)
super().__init__(
model=model,
sampler=sampler,
objective=objective,
posterior_transform=posterior_transform,
X_pending=X_pending,
constraints=constraints,
eta=eta,
)
CachedCholeskyMCSamplerMixin.__init__(
self, model=model, cache_root=cache_root, sampler=sampler
)
if prune_baseline:
X_baseline = prune_inferior_points(
model=model,
X=X_baseline,
objective=objective,
posterior_transform=posterior_transform,
constraints=self._constraints,
marginalize_dim=marginalize_dim,
)
self.register_buffer("X_baseline", X_baseline)
# registering buffers for _get_samples_and_objectives in the next `if` block
self.register_buffer("baseline_samples", None)
self.register_buffer("baseline_obj", None)
if self._cache_root:
self.q_in = -1
# set baseline samples
with torch.no_grad(): # this is _get_samples_and_objectives(X_baseline)
posterior = self.model.posterior(
X_baseline, posterior_transform=self.posterior_transform
)
# Note: The root decomposition is cached in two different places. It
# may be confusing to have two different caches, but this is not
# trivial to change since each is needed for a different reason:
# - LinearOperator caching to `posterior.mvn` allows for reuse within
# this function, which may be helpful if the same root decomposition
# is produced by the calls to `self.base_sampler` and
# `self._cache_root_decomposition`.
# - self._baseline_L allows a root decomposition to be persisted outside
# this method.
baseline_samples = self.get_posterior_samples(posterior)
baseline_obj = self.objective(baseline_samples, X=X_baseline)
# We make a copy here because we will write an attribute `base_samples`
# to `self.base_sampler.base_samples`, and we don't want to mutate
# `self.sampler`.
self.base_sampler = deepcopy(self.sampler)
self.baseline_samples = baseline_samples
self.baseline_obj = baseline_obj
self.register_buffer(
"_baseline_best_f",
self._compute_best_feasible_objective(
samples=baseline_samples, obj=baseline_obj
), # `sample_shape x batch_shape`-dim
)
self._baseline_L = self._compute_root_decomposition(posterior=posterior)
[docs]
def compute_best_f(self, obj: Tensor) -> Tensor:
"""Computes the best (feasible) noisy objective value.
Args:
obj: `sample_shape x batch_shape x q`-dim Tensor of objectives in forward.
Returns:
A `sample_shape x batch_shape`-dim Tensor of best feasible objectives.
"""
if self._cache_root:
val = self._baseline_best_f
else:
val = self._compute_best_feasible_objective(
samples=self.baseline_samples, obj=self.baseline_obj
)
# ensuring shape, dtype, device compatibility with obj
n_sample_dims = len(self.sample_shape)
view_shape = torch.Size(
[
*val.shape[:n_sample_dims], # sample dimensions
*(1,) * (obj.ndim - val.ndim - 1), # pad to match obj, without `q`-dim
*val.shape[n_sample_dims:], # the rest
]
)
return val.view(view_shape).to(obj) # obj.shape[:-1], i.e. without `q`-dim`
def _sample_forward(self, obj: Tensor) -> Tensor:
"""Evaluate qNoisyExpectedImprovement per objective value in `obj`.
Args:
obj: A `sample_shape x batch_shape x q`-dim Tensor of MC objective values.
Returns:
A `sample_shape x batch_shape x q`-dim Tensor of noisy improvement values.
"""
return (obj - self.compute_best_f(obj).unsqueeze(-1)).clamp_min(0)
def _get_samples_and_objectives(self, X: Tensor) -> tuple[Tensor, Tensor]:
r"""Compute samples at new points, using the cached root decomposition.
Args:
X: A `batch_shape x q x d`-dim tensor of inputs.
Returns:
A two-tuple `(samples, obj)`, where `samples` is a tensor of posterior
samples with shape `sample_shape x batch_shape x q x m`, and `obj` is a
tensor of MC objective values with shape `sample_shape x batch_shape x q`.
"""
q = X.shape[-2]
X_full = torch.cat([match_batch_shape(self.X_baseline, X), X], dim=-2)
# TODO: Implement more efficient way to compute posterior over both training and
# test points in GPyTorch (https://github.com/cornellius-gp/gpytorch/issues/567)
posterior = self.model.posterior(
X_full, posterior_transform=self.posterior_transform
)
if not self._cache_root:
samples_full = super().get_posterior_samples(posterior)
samples = samples_full[..., -q:, :]
obj_full = self.objective(samples_full, X=X_full)
# assigning baseline buffers so `best_f` can be computed in _sample_forward
self.baseline_obj, obj = obj_full[..., :-q], obj_full[..., -q:]
self.baseline_samples = samples_full[..., :-q, :]
else:
# handle one-to-many input transforms
n_plus_q = X_full.shape[-2]
n_w = posterior._extended_shape()[-2] // n_plus_q
q_in = q * n_w
self._set_sampler(q_in=q_in, posterior=posterior)
samples = self._get_f_X_samples(posterior=posterior, q_in=q_in)
obj = self.objective(samples, X=X_full[..., -q:, :])
return samples, obj
def _compute_best_feasible_objective(self, samples: Tensor, obj: Tensor) -> Tensor:
r"""Computes best feasible objective value from samples.
Args:
samples: `sample_shape x batch_shape x q x m`-dim posterior samples.
obj: A `sample_shape x batch_shape x q`-dim Tensor of MC objective values.
Returns:
A `sample_shape x batch_shape`-dim Tensor of best feasible objectives.
"""
return compute_best_feasible_objective(
samples=samples,
obj=obj,
constraints=self._constraints,
model=self.model,
objective=self.objective,
posterior_transform=self.posterior_transform,
X_baseline=self.X_baseline,
)
[docs]
class qProbabilityOfImprovement(SampleReducingMCAcquisitionFunction):
r"""MC-based batch Probability of Improvement.
Estimates the probability of improvement over the current best observed
value by sampling from the joint posterior distribution of the q-batch.
MC-based estimates of a probability involves taking expectation of an
indicator function; to support auto-differentiation, the indicator is
replaced with a sigmoid function with temperature parameter `tau`.
`qPI(X) = P(max Y >= best_f), Y ~ f(X), X = (x_1,...,x_q)`
Example:
>>> model = SingleTaskGP(train_X, train_Y)
>>> best_f = train_Y.max()[0]
>>> sampler = SobolQMCNormalSampler(1024)
>>> qPI = qProbabilityOfImprovement(model, best_f, sampler)
>>> qpi = qPI(test_X)
"""
def __init__(
self,
model: Model,
best_f: float | Tensor,
sampler: MCSampler | None = None,
objective: MCAcquisitionObjective | None = None,
posterior_transform: PosteriorTransform | None = None,
X_pending: Tensor | None = None,
tau: float = 1e-3,
constraints: list[Callable[[Tensor], Tensor]] | None = None,
eta: Tensor | float = 1e-3,
) -> None:
r"""q-Probability of Improvement.
Args:
model: A fitted model.
best_f: The best objective value observed so far (assumed noiseless). Can
be a `batch_shape`-shaped tensor, which in case of a batched model
specifies potentially different values for each element of the batch.
sampler: The sampler used to draw base samples. See `MCAcquisitionFunction`
more details.
objective: The MCAcquisitionObjective under which the samples are
evaluated. Defaults to `IdentityMCObjective()`.
NOTE: `ConstrainedMCObjective` for outcome constraints is deprecated in
favor of passing the `constraints` directly to this constructor.
posterior_transform: A PosteriorTransform (optional).
X_pending: A `m x d`-dim Tensor of `m` design points that have
points that have been submitted for function evaluation
but have not yet been evaluated. Concatenated into X upon
forward call. Copied and set to have no gradient.
tau: The temperature parameter used in the sigmoid approximation
of the step function. Smaller values yield more accurate
approximations of the function, but result in gradients
estimates with higher variance.
constraints: A list of constraint callables which map posterior samples to
a scalar. The associated constraint is considered satisfied if this
scalar is less than zero.
eta: Temperature parameter(s) governing the smoothness of the sigmoid
approximation to the constraint indicators. For more details, on this
parameter, see the docs of `compute_smoothed_feasibility_indicator`.
"""
super().__init__(
model=model,
sampler=sampler,
objective=objective,
posterior_transform=posterior_transform,
X_pending=X_pending,
constraints=constraints,
eta=eta,
)
best_f = torch.as_tensor(best_f, dtype=float).unsqueeze(-1) # adding batch dim
self.register_buffer("best_f", best_f)
self.register_buffer("tau", torch.as_tensor(tau, dtype=float))
def _sample_forward(self, obj: Tensor) -> Tensor:
r"""Evaluate qProbabilityOfImprovement per sample on the candidate set `X`.
Args:
obj: A `sample_shape x batch_shape x q`-dim Tensor of MC objective values.
Returns:
A `sample_shape x batch_shape x q`-dim Tensor of improvement indicators.
"""
improvement = obj - self.best_f.to(obj)
return torch.sigmoid(improvement / self.tau)
[docs]
class qSimpleRegret(SampleReducingMCAcquisitionFunction):
r"""MC-based batch Simple Regret.
Samples from the joint posterior over the q-batch and computes the simple regret.
`qSR(X) = E(max Y), Y ~ f(X), X = (x_1,...,x_q)`
Constraints should be provided as a `ConstrainedMCObjective`.
Passing `constraints` as an argument is not supported. This is because
`SampleReducingMCAcquisitionFunction` computes the acquisition values on the sample
level and then weights the sample-level acquisition values by a soft feasibility
indicator. Hence, it expects non-log acquisition function values to be
non-negative. `qSimpleRegret` acquisition values can be negative, so we instead use
a `ConstrainedMCObjective` which applies constraints to the objectives (e.g. before
computing the acquisition function) and shifts negative objective values using
by an infeasible cost to ensure non-negativity (before applying constraints and
shifting them back).
Example:
>>> model = SingleTaskGP(train_X, train_Y)
>>> sampler = SobolQMCNormalSampler(1024)
>>> qSR = qSimpleRegret(model, sampler)
>>> qsr = qSR(test_X)
"""
def __init__(
self,
model: Model,
sampler: MCSampler | None = None,
objective: MCAcquisitionObjective | None = None,
posterior_transform: PosteriorTransform | None = None,
X_pending: Tensor | None = None,
) -> None:
r"""q-Simple Regret.
Args:
model: A fitted model.
sampler: The sampler used to draw base samples. See `MCAcquisitionFunction`
more details.
objective: The MCAcquisitionObjective under which the samples are
evaluated. Defaults to `IdentityMCObjective()`.
posterior_transform: A PosteriorTransform (optional).
X_pending: A `m x d`-dim Tensor of `m` design points that have
points that have been submitted for function evaluation
but have not yet been evaluated. Concatenated into X upon
forward call. Copied and set to have no gradient.
"""
super().__init__(
model=model,
sampler=sampler,
objective=objective,
posterior_transform=posterior_transform,
X_pending=X_pending,
)
def _sample_forward(self, obj: Tensor) -> Tensor:
r"""Evaluate qSimpleRegret per sample on the candidate set `X`.
Args:
obj: A `sample_shape x batch_shape x q`-dim Tensor of MC objective values.
Returns:
A `sample_shape x batch_shape x q`-dim Tensor of simple regret values.
"""
return obj
[docs]
class qUpperConfidenceBound(SampleReducingMCAcquisitionFunction):
r"""MC-based batch Upper Confidence Bound.
Uses a reparameterization to extend UCB to qUCB for q > 1 (See Appendix A
of [Wilson2017reparam].)
`qUCB = E(max(mu + |Y_tilde - mu|))`, where `Y_tilde ~ N(mu, beta pi/2 Sigma)`
and `f(X)` has distribution `N(mu, Sigma)`.
Constraints should be provided as a `ConstrainedMCObjective`.
Passing `constraints` as an argument is not supported. This is because
`SampleReducingMCAcquisitionFunction` computes the acquisition values on the sample
level and then weights the sample-level acquisition values by a soft feasibility
indicator. Hence, it expects non-log acquisition function values to be
non-negative. `qSimpleRegret` acquisition values can be negative, so we instead use
a `ConstrainedMCObjective` which applies constraints to the objectives (e.g. before
computing the acquisition function) and shifts negative objective values using
by an infeasible cost to ensure non-negativity (before applying constraints and
shifting them back).
Example:
>>> model = SingleTaskGP(train_X, train_Y)
>>> sampler = SobolQMCNormalSampler(1024)
>>> qUCB = qUpperConfidenceBound(model, 0.1, sampler)
>>> qucb = qUCB(test_X)
"""
def __init__(
self,
model: Model,
beta: float,
sampler: MCSampler | None = None,
objective: MCAcquisitionObjective | None = None,
posterior_transform: PosteriorTransform | None = None,
X_pending: Tensor | None = None,
) -> None:
r"""q-Upper Confidence Bound.
Args:
model: A fitted model.
beta: Controls tradeoff between mean and standard deviation in UCB.
sampler: The sampler used to draw base samples. See `MCAcquisitionFunction`
more details.
objective: The MCAcquisitionObjective under which the samples are
evaluated. Defaults to `IdentityMCObjective()`.
posterior_transform: A PosteriorTransform (optional).
X_pending: A `batch_shape x m x d`-dim Tensor of `m` design points that have
points that have been submitted for function evaluation but have not yet
been evaluated. Concatenated into X upon forward call. Copied and set to
have no gradient.
"""
super().__init__(
model=model,
sampler=sampler,
objective=objective,
posterior_transform=posterior_transform,
X_pending=X_pending,
)
self.beta_prime = self._get_beta_prime(beta=beta)
def _get_beta_prime(self, beta: float) -> float:
return math.sqrt(beta * math.pi / 2)
def _sample_forward(self, obj: Tensor) -> Tensor:
r"""Evaluate qUpperConfidenceBound per sample on the candidate set `X`.
Args:
obj: A `sample_shape x batch_shape x q`-dim Tensor of MC objective values.
Returns:
A `sample_shape x batch_shape x q`-dim Tensor of acquisition values.
"""
mean = obj.mean(dim=0)
return mean + self.beta_prime * (obj - mean).abs()
[docs]
class qLowerConfidenceBound(qUpperConfidenceBound):
r"""MC-based batched lower confidence bound.
This acquisition function is useful for confident/risk-averse decision making.
This acquisition function is intended to be maximized as with qUpperConfidenceBound,
but the qLowerConfidenceBound will be pessimistic in the face of uncertainty and
lead to conservative candidates.
"""
def _get_beta_prime(self, beta: float) -> float:
"""Multiply beta prime by -1 to get the lower confidence bound."""
return -super()._get_beta_prime(beta=beta)
