Models play an essential role in Bayesian Optimization (BO). A model is used as a
surrogate function for the actual underlying black box function to be optimized.
In BoTorch, a
Model maps a set of design points to a posterior probability
distribution of its output(s) over the design points.
In BO, the model used is traditionally a Gaussian Process (GP),
in which case the posterior distribution, by definition is a multivariate
normal. However, with the exception of some of the analytic acquisition
functions in the
module, BoTorch makes no assumption on the model being a GP, or on the
posterior being a multivariate normal. The only requirement for using
BoTorch's Monte-Carlo based acquisition functions is that the model returns a
Posterior object that implements an
rsample() method for sampling from the posterior of the model. If you wish to
use gradient-based optimization algorithms, the model should allow
back-propagating gradients through the samples to the model input.
The BoTorch Model Interface
BoTorch models are PyTorch modules that implement the light-weight
Model interface. A BoTorch
Model requires only
posterior() method that takes in a Tensor
X of design points,
and returns a
Posterior object describing
the (joint) probability distribution of the model output(s) over the design
When working with GPs,
provides a base class for conveniently wrapping GPyTorch models.
Models may have multiple outputs, multiple inputs, and may exploit correlation between different inputs. BoTorch uses the following terminology to distinguish these model types:
- Multi-Output Model: a
Model(as in the BoTorch object) with multiple outputs.
- Multi-Task Model: a
Modelmaking use of a logical grouping of inputs/observations (as in the underlying process). For example, there could be multiple tasks where each task has a different fidelity.
Note the following:
- A multi-task (MT) model may or may not be a multi-output model.
- Conversely, a multi-output (MO) model may or may not be a multi-task model.
- If a model is both, we refer to it as a multi-task-multi-output (MTMO) model.
Standard BoTorch Models
BoTorch provides several GPyTorch models to cover most standard BO use cases:
These models use the same training data for all outputs and assume conditional
independence of the outputs given the input. If different training data is
required for each output, use a
SingleTaskGP: a single-task exact GP that infers a homoskedastic noise level (no noise observations).
FixedNoiseGP: a single-task exact GP that uses fixed observation noise levels (requires noise observations).
HeteroskedasticSingleTaskGP: a single-task exact GP that models heteroskedastic noise using an additional internal GP model (requires noise observations).
SaasFullyBayesianSingleTaskGP: a fully Bayesian single-task GP with the SAAS prior. This model is suitable for sample-efficient high-dimensional Bayesian optimization.
Model List of Single-Task GPs
ModelListGP: A multi-output model in which outcomes are modeled independently, given a list of any type of single-task GP. This model should be used when the same training data is not used for all outputs.
MultiTaskGP: a Hadamard multi-task, multi-output GP using an ICM kernel, inferring the noise level (does not require noise observations).
FixedNoiseMultiTaskGP: a Hadamard multi-task, multi-output GP using an ICM kernel, with fixed observation noise levels (requires noise observations).
All of the above models use Matérn 5/2 kernels with Automatic Relevance Discovery (ARD), and have reasonable priors on hyperparameters that make them work well in settings where the input features are normalized to the unit cube and the observations are standardized (zero mean, unit variance).
Implementing Custom Models
The configurability of the above models is limited (for instance, it is not straightforward to use a different kernel). Doing so is an intentional design decision -- we believe that having a few simple and easy-to-understand models for basic use cases is more valuable than having a highly complex and configurable model class whose implementation is difficult to understand.
Instead, we advocate that users implement their own models to cover more specialized use cases. The light-weight nature of BoTorch's Model API makes this easy to do. See the Using a custom BoTorch model in Ax tutorial for an example.
If you happen to implement a model that would be useful for other researchers as well (and involves more than just swapping out the Matérn kernel for an RBF kernel), please consider contributing this model to BoTorch.