HEBO (Heteroscedastic and Evolutionary Bayesian Optimisation)

Class or Function Names

• HEBOSampler

Installation

# Install the dependencies.
pip install optunahub hebo

# NOTE: Below is optional, but pymoo must be installed after NumPy for faster HEBOSampler,
# we run the following command to make sure that the compiled version is installed.
pip install --upgrade pymoo

APIs

• HEBOSampler(search_space: dict[str, BaseDistribution] | None = None, *, seed: int | None = None, constant_liar: bool = False, independent_sampler: BaseSampler | None = None)

  • search_space: By specifying search_space, the sampling speed at each iteration becomes slightly quicker, but this argument is not necessary to run this sampler.

    Example:

    search_space = {
        "x": optuna.distributions.FloatDistribution(-5, 5),
        "y": optuna.distributions.FloatDistribution(-5, 5),
    }
    HEBOSampler(search_space=search_space)
    

  • seed: A seed for the initialization of HEBOSampler. Please note that the Bayesian optimization part is not deterministic even if seed is fixed due to the backend implementation.

  • constant_liar: If True, penalize running trials to avoid suggesting parameter configurations nearby. Default is False.

    • Note: Abnormally terminated trials often leave behind a record with a state of RUNNING in the storage. Such “zombie” trial parameters will be avoided by the constant liar algorithm during subsequent sampling. When using an optuna.storages.RDBStorage, it is possible to enable the heartbeat_interval to change the records for abnormally terminated trials to FAIL. (This note is quoted from TPESampler.)
    • Note: It is recommended to set this value to True during distributed optimization to avoid having multiple workers evaluating similar parameter configurations. In particular, if each objective function evaluation is costly and the durations of the running states are significant, and/or the number of workers is high. (This note is quoted from TPESampler.)
    • Note: HEBO algorithm involves multi-objective optimization of multiple acquisition functions. While constant_liar is a simple way to get diverse params for parallel optimization, it may not be the best approach for HEBO.

  • independent_sampler: A optuna.samplers.BaseSampler instance that is used for independent sampling. The parameters not contained in the relative search space are sampled by this sampler. If None is specified, optuna.samplers.RandomSampler is used as the default.

Example

import optuna
import optunahub


def objective(trial: optuna.trial.Trial) -> float:
    x = trial.suggest_float("x", -10, 10)
    y = trial.suggest_int("y", -10, 10)
    return x**2 + y**2


module = optunahub.load_module("samplers/hebo")
sampler = module.HEBOSampler()
study = optuna.create_study(sampler=sampler)
study.optimize(objective, n_trials=100)

print(study.best_trial.params, study.best_trial.value)

See example.py for a full example.

Note that it may slightly speed up the sampling routine by giving the search space directly to HEBOSampler since Optuna can skip the search space inference. For example, the instantiation of HEBOSampler above can be modified as follows:

search_space={
    "x": optuna.distributions.FloatDistribution(-10, 10),
    "y": optuna.distributions.IntDistribution(-10, 10),
}
sampler = module.HEBOSampler(search_space=search_space)

However, users need to make sure that the provided search space and the search space defined in the objective function must be consistent.

Others

HEBO is the winning submission to the NeurIPS 2020 Black-Box Optimisation Challenge. Please refer to the official repository of HEBO for more details.

Reference

Cowen-Rivers, Alexander I., et al. “An Empirical Study of Assumptions in Bayesian Optimisation.” arXiv preprint arXiv:2012.03826 (2021).