BayesianOptimization

Functionality for Bayesian optimization, the data-efficient optimization of objective functions.
Bayesian optimization recommends new points at which to query your objective by placing a
probabilistic model over historic data then, typically, optimizing an _acquisition function_
which quantifies how useful it would be to evaluate the objective at any given set of points.

Reexports

Definitions

data TagStream : Type -> Type

  A `Stream`-like collection where each successive element can extend the set of `Tag`s.

Totality: total
Visibility: public export
Constructor: 

(::) : a -> Inf (Tag (TagStream a)) -> TagStream a

take : (n : Nat) -> TagStream a -> Tag (Vect n a)

  Take `n` values from a `TagStream`, sequencing the `Tag` effects.

Totality: total
Visibility: public export

iterate : (a -> Tag a) -> a -> Tag (TagStream a)

  Create an infinite stream of values from a generator function and a starting value.

Visibility: export

step : (Tensor (n :: features) F64 -> Tag (Tensor (n :: targets) F64)) -> {auto probabilisticModel : ProbabilisticModel features targets marginal model} -> (Dataset features targets -> model -> Tag model) -> ReaderT (DataModel model) Tag (Tensor (1 :: features) F64) -> DataModel model -> Tag (DataModel model)

  Construct a single simple Bayesian optimization step.
  
  @objective The objective function to optimize.
  @train Used to train the model on new data.
  @tactic The tactic, such as an optimized acquisition function, to find a new point from the
    data and model

Totality: total
Visibility: export