def observed_hypervolume(
modelbridge: modelbridge_module.array.ArrayModelBridge,
objective_thresholds: Optional[TRefPoint] = None,
optimization_config: Optional[MultiObjectiveOptimizationConfig] = None,
) -> float:
"""Calculate hypervolume of a pareto frontier based on observed data.
Given observed data, return the hypervolume of the pareto frontier formed from
those outcomes.
Args:
modelbridge: Modelbridge that holds previous training data.
objective_thresholds: point defining the origin of hyperrectangles that
can contribute to hypervolume.
observation_features: observation features to predict. Model's training
data used by default if unspecified.
optimization_config: Optimization config
Returns:
(float) calculated hypervolume.
"""
# Get observation_data from current training data.
observation_data = [obs.data for obs in modelbridge.get_training_data()]
observation_features = [
obs.features for obs in modelbridge.get_training_data()
]
return hypervolume(
modelbridge=modelbridge,
objective_thresholds=objective_thresholds,
observation_features=observation_features,
observation_data=observation_data,
optimization_config=optimization_config,
use_model_predictions=False,
)
def observed_pareto_frontier(
modelbridge: modelbridge_module.array.ArrayModelBridge,
objective_thresholds: Optional[TRefPoint] = None,
optimization_config: Optional[MultiObjectiveOptimizationConfig] = None,
) -> List[ObservationData]:
"""Generate a pareto frontier based on observed data.
Given observed data, return those outcomes in the pareto frontier.
Args:
modelbridge: Modelbridge that holds previous training data.
objective_thresholds: metric values bounding the region of interest in
the objective outcome space.
optimization_config: Optimization config
Returns:
Data representing points on the pareto frontier.
"""
# Get observation_data from current training data
observation_data = [obs.data for obs in modelbridge.get_training_data()]
return pareto_frontier(
modelbridge=modelbridge,
objective_thresholds=objective_thresholds,
observation_data=observation_data,
optimization_config=optimization_config,
)
def _get_modelbridge_training_data(
modelbridge: modelbridge_module.array.ArrayModelBridge,
) -> Tuple[List[ObservationFeatures], List[ObservationData],
List[Optional[str]]]:
obs = modelbridge.get_training_data()
obs_feats, obs_data, arm_names = [], [], []
for ob in obs:
obs_feats.append(ob.features)
obs_data.append(ob.data)
arm_names.append(ob.arm_name)
return obs_feats, obs_data, arm_names
def predicted_pareto_frontier(
modelbridge: modelbridge_module.array.ArrayModelBridge,
objective_thresholds: Optional[TRefPoint] = None,
observation_features: Optional[List[ObservationFeatures]] = None,
optimization_config: Optional[MultiObjectiveOptimizationConfig] = None,
) -> List[Observation]:
"""Generate a pareto frontier based on the posterior means of given
observation features.
Given a model and features to evaluate use the model to predict which points
lie on the pareto frontier.
Args:
modelbridge: Modelbridge used to predict metrics outcomes.
objective_thresholds: metric values bounding the region of interest in
the objective outcome space.
observation_features: observation features to predict. Model's training
data used by default if unspecified.
optimization_config: Optimization config
Returns:
Observations representing points on the pareto frontier.
"""
if observation_features is None:
observation_features = []
arm_names = []
for obs in modelbridge.get_training_data():
observation_features.append(obs.features)
arm_names.append(obs.arm_name)
else:
arm_names = None
if not observation_features:
raise ValueError(
"Must receive observation_features as input or the model must "
"have training data."
)
pareto_observations = pareto_frontier(
modelbridge=modelbridge,
objective_thresholds=objective_thresholds,
observation_features=observation_features,
optimization_config=optimization_config,
arm_names=arm_names,
)
return pareto_observations
def predicted_hypervolume(
modelbridge: modelbridge_module.array.ArrayModelBridge,
objective_thresholds: Optional[TRefPoint] = None,
observation_features: Optional[List[ObservationFeatures]] = None,
optimization_config: Optional[MultiObjectiveOptimizationConfig] = None,
) -> float:
"""Calculate hypervolume of a pareto frontier based on the posterior means of
given observation features.
Given a model and features to evaluate calculate the hypervolume of the pareto
frontier formed from their predicted outcomes.
Args:
modelbridge: Modelbridge used to predict metrics outcomes.
objective_thresholds: point defining the origin of hyperrectangles that
can contribute to hypervolume.
observation_features: observation features to predict. Model's training
data used by default if unspecified.
optimization_config: Optimization config
Returns:
calculated hypervolume.
"""
observation_features = (
observation_features
if observation_features is not None
else [obs.features for obs in modelbridge.get_training_data()]
)
if not observation_features:
raise ValueError(
"Must receive observation_features as input or the model must "
"have training data."
)
return hypervolume(
modelbridge=modelbridge,
objective_thresholds=objective_thresholds,
observation_features=observation_features,
optimization_config=optimization_config,
)
def _get_modelbridge_training_data(
modelbridge: modelbridge_module.array.ArrayModelBridge,
) -> Tuple[List[ObservationFeatures], List[ObservationData],
List[Optional[str]]]:
obs = modelbridge.get_training_data()
return _unpack_observations(obs=obs)