import functools
import logging
import time
import warnings
import ConfigSpace as CS
import ConfigSpace.hyperparameters as csh
import numpy as np
import pandas as pd
import deephyper.core.exceptions
import deephyper.skopt
from deephyper.problem._hyperparameter import convert_to_skopt_space
from deephyper.search._search import Search
from deephyper.skopt.moo import non_dominated_set, non_dominated_set_ranked
from sklearn.ensemble import GradientBoostingRegressor
from sklearn.base import is_regressor
from deephyper.skopt.utils import use_named_args
# Adapt minimization -> maximization with DeepHyper
MAP_multi_point_strategy = {"cl_min": "cl_max", "cl_max": "cl_min", "qUCB": "qLCB"}
MAP_acq_func = {"UCB": "LCB", "qUCB": "qLCB"}
MAP_filter_failures = {"min": "max"}
# schedulers
def scheduler_periodic_exponential_decay(i, eta_0, periode, rate, delay):
"""Periodic exponential decay scheduler for exploration-exploitation."""
eta_i = eta_0 * np.exp(-rate * ((i - 1 - delay) % periode))
return eta_i
[docs]class CBO(Search):
"""Centralized Bayesian Optimisation Search, previously named as "Asynchronous Model-Based Search" (AMBS). It follows a manager-workers architecture where the manager runs the Bayesian optimization loop and workers execute parallel evaluations of the black-box function.
Example Usage:
>>> search = CBO(problem, evaluator)
>>> results = search.search(max_evals=100, timeout=120)
Args:
problem (HpProblem): Hyperparameter problem describing the search space to explore.
evaluator (Evaluator): An ``Evaluator`` instance responsible of distributing the tasks.
random_state (int, optional): Random seed. Defaults to ``None``.
log_dir (str, optional): Log directory where search's results are saved. Defaults to ``"."``.
verbose (int, optional): Indicate the verbosity level of the search. Defaults to ``0``.
surrogate_model (Union[str,sklearn.base.RegressorMixin], optional): Surrogate model used by the Bayesian optimization. Can be a value in ``["RF", "GP", "ET", "GBRT", "DUMMY"]`` or a sklearn regressor. ``"RF"`` is for Random-Forest which is the best compromise between speed and quality when performing a lot of parallel evaluations, i.e., reaching more than hundreds of evaluations. ``"GP"`` is for Gaussian-Process which is the best choice when maximizing the quality of iteration but quickly slow down when reaching hundreds of evaluations, also it does not support conditional search space. ``"ET"`` is for Extra-Tree, faster than random forest but with worse mean estimate and poor uncertainty quantification capabilities. ``"GBRT"`` is for Gradient-Boosting Regression Tree, it has better mean estimate than other tree-based method worse uncertainty quantification capabilities and slower than ``"RF"``. Defaults to ``"RF"``.
acq_func (str, optional): Acquisition function used by the Bayesian optimization. Can be a value in ``["UCB", "EI", "PI", "gp_hedge"]``. Defaults to ``"UCB"``.
acq_optimizer (str, optional): Method used to minimze the acquisition function. Can be a value in ``["sampling", "lbfgs"]``. Defaults to ``"auto"``.
kappa (float, optional): Manage the exploration/exploitation tradeoff for the "UCB" acquisition function. Defaults to ``1.96`` which corresponds to 95% of the confidence interval.
xi (float, optional): Manage the exploration/exploitation tradeoff of ``"EI"`` and ``"PI"`` acquisition function. Defaults to ``0.001``.
n_points (int, optional): The number of configurations sampled from the search space to infer each batch of new evaluated configurations.
filter_duplicated (bool, optional): Force the optimizer to sample unique points until the search space is "exhausted" in the sens that no new unique points can be found given the sampling size ``n_points``. Defaults to ``True``.
multi_point_strategy (str, optional): Definition of the constant value use for the Liar strategy. Can be a value in ``["cl_min", "cl_mean", "cl_max", "qUCB"]``. All ``"cl_..."`` strategies follow the constant-liar scheme, where if $N$ new points are requested, the surrogate model is re-fitted $N-1$ times with lies (respectively, the minimum, mean and maximum objective found so far; for multiple objectives, these are the minimum, mean and maximum of the individual objectives) to infer the acquisition function. Constant-Liar strategy have poor scalability because of this repeated re-fitting. The ``"qUCB"`` strategy is much more efficient by sampling a new $kappa$ value for each new requested point without re-fitting the model, but it is only compatible with ``acq_func == "UCB"``. Defaults to ``"cl_max"``.
n_jobs (int, optional): Number of parallel processes used to fit the surrogate model of the Bayesian optimization. A value of ``-1`` will use all available cores. Not used in ``surrogate_model`` if passed as own sklearn regressor. Defaults to ``1``.
n_initial_points (int, optional): Number of collected objectives required before fitting the surrogate-model. Defaults to ``10``.
initial_point_generator (str, optional): Sets an initial points generator. Can be either ``["random", "sobol", "halton", "hammersly", "lhs", "grid"]``. Defaults to ``"random"``.
initial_points (List[Dict], optional): A list of initial points to evaluate where each point is a dictionnary where keys are names of hyperparameters and values their corresponding choice. Defaults to ``None`` for them to be generated randomly from the search space.
sync_communcation (bool, optional): Performs the search in a batch-synchronous manner. Defaults to ``False`` for asynchronous updates.
filter_failures (str, optional): Replace objective of failed configurations by ``"min"`` or ``"mean"``. If ``"ignore"`` is passed then failed configurations will be filtered-out and not passed to the surrogate model. For multiple objectives, failure of any single objective will lead to treating that configuration as failed and each of these multiple objective will be replaced by their individual ``"min"`` or ``"mean"`` of past configurations. Defaults to ``"mean"`` to replace by failed configurations by the running mean of objectives.
max_failures (int, optional): Maximum number of failed configurations allowed before observing a valid objective value when ``filter_failures`` is not equal to ``"ignore"``. Defaults to ``100``.
moo_scalarization_strategy (str, optional): Scalarization strategy used in multiobjective optimization. Can be a value in ``["Linear", "Chebyshev", "AugChebyshev", "PBI", "Quadratic", "rLinear", "rChebyshev", "rAugChebyshev", "rPBI", "rQuadratic"]``. Defaults to ``"Chebyshev"``.
moo_scalarization_weight (list, optional): Scalarization weights to be used in multiobjective optimization with length equal to the number of objective functions. Defaults to ``None``.
scheduler (dict, callable, optional): a method to manage the the value of ``kappa, xi`` with iterations. Defaults to ``None`` which does not use any scheduler.
objective_scaler (str, optional): a way to map the objective space to some other support for example to normalize it. Defaults to ``"auto"`` which automatically set it to "identity" for any surrogate model except "RF" which will use "minmaxlog".
stopper (Stopper, optional): a stopper to leverage multi-fidelity when evaluating the function. Defaults to ``None`` which does not use any stopper.
"""
def __init__(
self,
problem,
evaluator,
random_state: int = None,
log_dir: str = ".",
verbose: int = 0,
surrogate_model="RF",
acq_func: str = "UCB",
acq_optimizer: str = "auto",
kappa: float = 1.96,
xi: float = 0.001,
n_points: int = 10000,
filter_duplicated: bool = True,
update_prior: bool = False,
multi_point_strategy: str = "cl_max",
n_jobs: int = 1, # 32 is good for Theta
n_initial_points: int = 10,
initial_point_generator: str = "random",
initial_points=None,
sync_communication: bool = False,
filter_failures: str = "mean",
max_failures: int = 100,
moo_scalarization_strategy: str = "Chebyshev",
moo_scalarization_weight=None,
scheduler=None,
objective_scaler="auto",
stopper=None,
**kwargs,
):
super().__init__(problem, evaluator, random_state, log_dir, verbose)
# get the __init__ parameters
self._init_params = locals()
# check input parameters
if not (type(n_jobs) is int):
raise ValueError(f"Parameter n_jobs={n_jobs} should be an integer value!")
surrogate_model_allowed = ["RF", "ET", "GBRT", "DUMMY", "GP", "MF"]
if surrogate_model in surrogate_model_allowed:
base_estimator = self._get_surrogate_model(
surrogate_model,
n_jobs,
random_state=self._random_state.randint(0, 2**32),
)
elif is_regressor(surrogate_model):
base_estimator = surrogate_model
else:
raise ValueError(
f"Parameter 'surrogate_model={surrogate_model}' should have a value in {surrogate_model_allowed}, or be a sklearn regressor!"
)
acq_func_allowed = ["UCB", "EI", "PI", "gp_hedge", "qUCB"]
if not (acq_func in acq_func_allowed):
raise ValueError(
f"Parameter 'acq_func={acq_func}' should have a value in {acq_func_allowed}!"
)
if not (np.isscalar(kappa)):
raise ValueError("Parameter 'kappa' should be a scalar value!")
if not (np.isscalar(xi)):
raise ValueError("Parameter 'xi' should be a scalar value!")
if not (type(n_points) is int):
raise ValueError("Parameter 'n_points' shoud be an integer value!")
if not (type(filter_duplicated) is bool):
raise ValueError(
f"Parameter filter_duplicated={filter_duplicated} should be a boolean value!"
)
if not (type(max_failures) is int):
raise ValueError(
f"Parameter max_failures={max_failures} should be an integer value!"
)
moo_scalarization_strategy_allowed = [
"Linear",
"Chebyshev",
"AugChebyshev",
"PBI",
"Quadratic",
]
moo_scalarization_strategy_allowed = moo_scalarization_strategy_allowed + [
f"r{s}" for s in moo_scalarization_strategy_allowed
]
if not (moo_scalarization_strategy in moo_scalarization_strategy_allowed):
raise ValueError(
f"Parameter 'moo_scalarization_strategy={acq_func}' should have a value in {moo_scalarization_strategy_allowed}!"
)
self._moo_scalarization_strategy = moo_scalarization_strategy
self._moo_scalarization_weight = moo_scalarization_weight
multi_point_strategy_allowed = [
"cl_min",
"cl_mean",
"cl_max",
"topk",
"boltzmann",
"qUCB",
]
if not (multi_point_strategy in multi_point_strategy_allowed):
raise ValueError(
f"Parameter multi_point_strategy={multi_point_strategy} should have a value in {multi_point_strategy_allowed}!"
)
self._n_initial_points = n_initial_points
self._initial_points = []
if initial_points is not None and len(initial_points) > 0:
for point in initial_points:
if isinstance(point, list):
self._initial_points.append(point)
elif isinstance(point, dict):
self._initial_points.append(
[point[hp_name] for hp_name in problem.hyperparameter_names]
)
else:
raise ValueError(
f"Initial points should be dict or list but {type(point)} was given!"
)
self._multi_point_strategy = MAP_multi_point_strategy.get(
multi_point_strategy, multi_point_strategy
)
self._fitted = False
# check if it is possible to convert the ConfigSpace to standard skopt Space
if (
isinstance(self._problem.space, CS.ConfigurationSpace)
and len(self._problem.space.get_forbiddens()) == 0
and len(self._problem.space.get_conditions()) == 0
):
self._opt_space = convert_to_skopt_space(
self._problem.space, surrogate_model=surrogate_model
)
else:
self._opt_space = self._problem.space
self._opt = None
self._opt_kwargs = dict(
dimensions=self._opt_space,
base_estimator=base_estimator,
# optimizer
initial_point_generator=initial_point_generator,
acq_optimizer=acq_optimizer,
acq_optimizer_kwargs={
"n_points": n_points,
"filter_duplicated": filter_duplicated,
"update_prior": update_prior,
"n_jobs": n_jobs,
"filter_failures": MAP_filter_failures.get(
filter_failures, filter_failures
),
"max_failures": max_failures,
"boltzmann_gamma": 1,
},
# acquisition function
acq_func=MAP_acq_func.get(acq_func, acq_func),
acq_func_kwargs={"xi": xi, "kappa": kappa},
n_initial_points=self._n_initial_points,
initial_points=self._initial_points,
random_state=self._random_state,
moo_scalarization_strategy=self._moo_scalarization_strategy,
moo_scalarization_weight=self._moo_scalarization_weight,
objective_scaler=objective_scaler,
)
self._gather_type = "ALL" if sync_communication else "BATCH"
# scheduler policy
self.scheduler = None
if type(scheduler) is dict:
scheduler = scheduler.copy()
scheduler_type = scheduler.pop("type", None)
assert scheduler_type in ["periodic-exp-decay"]
if scheduler_type == "periodic-exp-decay":
scheduler_params = {
"periode": 25,
"rate": 0.1,
"delay": n_initial_points,
}
scheduler_func = scheduler_periodic_exponential_decay
scheduler_params.update(scheduler)
eta_0 = np.array([kappa, xi])
self.scheduler = functools.partial(
scheduler_func, eta_0=eta_0, **scheduler_params
)
logging.info(
f"Set up scheduler '{scheduler_type}' with parameters '{scheduler_params}'"
)
# stopper
self._evaluator._stopper = stopper
def _setup_optimizer(self):
if self._fitted:
self._opt_kwargs["n_initial_points"] = 0
self._opt = deephyper.skopt.Optimizer(**self._opt_kwargs)
def _apply_scheduler(self, i):
"""Apply scheduler policy and update corresponding values in Optimizer."""
if self.scheduler is not None:
kappa, xi = self.scheduler(i)
values = {"kappa": kappa, "xi": xi}
logging.info(
f"Updated exploration-exploitation policy with {values} from scheduler"
)
self._opt.acq_func_kwargs.update(values)
def _search(self, max_evals, timeout):
if self._opt is None:
self._setup_optimizer()
num_evals_done = 0
num_local_evals_done = 0
logging.info(f"Asking {self._evaluator.num_workers} initial configurations...")
t1 = time.time()
new_X = self._opt.ask(n_points=self._evaluator.num_workers)
logging.info(f"Asking took {time.time() - t1:.4f} sec.")
# Transform list to dict configurations
logging.info("Transforming configurations to dict...")
t1 = time.time()
new_batch = []
for x in new_X:
new_cfg = self._to_dict(x)
new_batch.append(new_cfg)
logging.info(f"Transformation took {time.time() - t1:.4f} sec.")
# submit new configurations
logging.info(f"Submitting {len(new_batch)} configurations...")
t1 = time.time()
self._evaluator.submit(new_batch)
logging.info(f"Submition took {time.time() - t1:.4f} sec.")
# Main loop
while max_evals < 0 or num_evals_done < max_evals:
# Collecting finished evaluations
logging.info("Gathering jobs...")
t1 = time.time()
new_results = self._evaluator.gather(self._gather_type, size=1)
if isinstance(new_results, tuple) and len(new_results) == 2:
local_results, other_results = new_results
new_results = local_results + other_results
num_new_local_results = len(local_results)
num_new_other_results = len(other_results)
logging.info(
f"Gathered {num_new_local_results} local job(s) and {num_new_other_results} other job(s) in {time.time() - t1:.4f} sec."
)
else:
num_new_local_results = len(new_results)
logging.info(
f"Gathered {num_new_local_results} job(s) in {time.time() - t1:.4f} sec."
)
num_local_evals_done += num_new_local_results
if num_new_local_results > 0:
logging.info("Dumping evaluations...")
t1 = time.time()
self._evaluator.dump_evals(log_dir=self._log_dir)
logging.info(f"Dumping took {time.time() - t1:.4f} sec.")
num_evals_done += len(new_results)
if max_evals > 0 and num_evals_done >= max_evals:
break
# Transform configurations to list to fit optimizer
logging.info("Transforming received configurations to list...")
t1 = time.time()
opt_X = [] # input configuration
opt_y = [] # objective value
opt_b = [] # budget (optional)
# for cfg, obj in new_results:
for job_i in new_results:
cfg, obj = job_i
x = list(cfg.values())
# retrieve budget consumed by job with multiple observations
if job_i.observations is not None:
# # TODO: use ALC to reduce the problem to a scalar maximization/estimation
from deephyper.stopper._lcmodel_stopper import (
area_learning_curve,
)
# z_values: are the steps z from the budget function b(z)
# the job observations returns the observed (budgets, objectives)
# steps can be deduced from the length of these lists
# y_values: are the objective f(b(z))
_, y_values = np.array(job_i.observations)
z_values = np.arange(len(y_values)) + 1
y_values = -y_values
y = area_learning_curve(
z_values, y_values, z_max=self._evaluator._stopper.max_steps
)
opt_X.append(x)
opt_y.append(y)
# TODO: the following approach will not scale!
# for b, y in zip(*job_i.observations):
# opt_X.append(x)
# opt_b.append(b)
# opt_y.append(-y)
# single observation returned without budget
else:
if np.all(np.isreal(obj)):
opt_X.append(x)
opt_y.append(np.negative(obj).tolist()) # !maximizing
elif (type(obj) is str and "F" == obj[0]) or np.any(
type(objval) is str and "F" == objval[0] for objval in obj
):
if (
self._opt_kwargs["acq_optimizer_kwargs"][
"filter_failures"
]
== "ignore"
):
continue
else:
opt_X.append(x)
opt_y.append("F")
logging.info(f"Transformation took {time.time() - t1:.4f} sec.")
# apply scheduler
self._apply_scheduler(i=num_local_evals_done)
logging.info("Fitting the optimizer...")
t1 = time.time()
if len(opt_y) > 0:
opt_b = None if len(opt_b) == 0 else opt_b
self._opt.tell(opt_X, opt_y, budget=opt_b)
logging.info(f"Fitting took {time.time() - t1:.4f} sec.")
logging.info(f"Asking {num_new_local_results} new configurations...")
t1 = time.time()
new_X = self._opt.ask(
n_points=num_new_local_results, strategy=self._multi_point_strategy
)
logging.info(f"Asking took {time.time() - t1:.4f} sec.")
# Transform list to dict configurations
logging.info("Transforming configurations to dict...")
t1 = time.time()
new_batch = []
for x in new_X:
new_cfg = self._to_dict(x)
new_batch.append(new_cfg)
logging.info(f"Transformation took {time.time() - t1:.4f} sec.")
# submit new configurations
logging.info(f"Submitting {len(new_batch)} configurations...")
t1 = time.time()
self._evaluator.submit(new_batch)
logging.info(f"Submition took {time.time() - t1:.4f} sec.")
def _get_surrogate_model(
self, name: str, n_jobs: int = None, random_state: int = None
):
"""Get a surrogate model from Scikit-Optimize.
Args:
name (str): name of the surrogate model.
n_jobs (int): number of parallel processes to distribute the computation of the surrogate model.
Raises:
ValueError: when the name of the surrogate model is unknown.
"""
accepted_names = ["RF", "ET", "GBRT", "DUMMY", "GP", "MF"]
if not (name in accepted_names):
raise ValueError(
f"Unknown surrogate model {name}, please choose among {accepted_names}."
)
if name == "RF":
surrogate = deephyper.skopt.learning.RandomForestRegressor(
# n_estimators=100,
# max_features=1,
# min_samples_leaf=3,
n_jobs=n_jobs,
random_state=random_state,
)
elif name == "ET":
surrogate = deephyper.skopt.learning.ExtraTreesRegressor(
# n_estimators=100,
# min_samples_leaf=3,
n_jobs=n_jobs,
random_state=random_state,
)
elif name == "GBRT":
gbrt = GradientBoostingRegressor(n_estimators=30, loss="quantile")
surrogate = deephyper.skopt.learning.GradientBoostingQuantileRegressor(
base_estimator=gbrt, n_jobs=n_jobs, random_state=random_state
)
elif name == "MF":
try:
surrogate = deephyper.skopt.learning.MondrianForestRegressor(
n_estimators=100, n_jobs=n_jobs, random_state=random_state
)
except AttributeError:
raise deephyper.core.exceptions.MissingRequirementError(
"Installing 'deephyper/scikit-garden' is required to use MondrianForest (MF) regressor as a surrogate model!"
)
else: # for DUMMY and GP
surrogate = name
return surrogate
def _return_cond(self, cond, cst_new):
parent = cst_new.get_hyperparameter(cond.parent.name)
child = cst_new.get_hyperparameter(cond.child.name)
if type(cond) == CS.EqualsCondition:
value = cond.value
cond_new = CS.EqualsCondition(child, parent, cond.value)
elif type(cond) == CS.GreaterThanCondition:
value = cond.value
cond_new = CS.GreaterThanCondition(child, parent, value)
elif type(cond) == CS.NotEqualsCondition:
value = cond.value
cond_new = CS.GreaterThanCondition(child, parent, value)
elif type(cond) == CS.LessThanCondition:
value = cond.value
cond_new = CS.GreaterThanCondition(child, parent, value)
elif type(cond) == CS.InCondition:
values = cond.values
cond_new = CS.GreaterThanCondition(child, parent, values)
else:
print("Not supported type" + str(type(cond)))
return cond_new
def _return_forbid(self, cond, cst_new):
if type(cond) == CS.ForbiddenEqualsClause or type(cond) == CS.ForbiddenInClause:
hp = cst_new.get_hyperparameter(cond.hyperparameter.name)
if type(cond) == CS.ForbiddenEqualsClause:
value = cond.value
cond_new = CS.ForbiddenEqualsClause(hp, value)
elif type(cond) == CS.ForbiddenInClause:
values = cond.values
cond_new = CS.ForbiddenInClause(hp, values)
else:
print("Not supported type" + str(type(cond)))
return cond_new
[docs] def fit_surrogate(self, df):
"""Fit the surrogate model of the search from a checkpointed Dataframe.
Args:
df (str|DataFrame): a checkpoint from a previous search.
Example Usage:
>>> search = CBO(problem, evaluator)
>>> search.fit_surrogate("results.csv")
"""
if type(df) is str and df[-4:] == ".csv":
df = pd.read_csv(df)
assert isinstance(df, pd.DataFrame)
self._fitted = True
if self._opt is None:
self._setup_optimizer()
hp_names = [f"p:{name}" for name in self._problem.hyperparameter_names]
try:
x = df[hp_names].values.tolist()
# check single or multiple objectives
if "objective" in df.columns:
y = df.objective.tolist()
else:
y = df.filter(regex=r"^objective_\d+$").values.tolist()
except KeyError:
raise ValueError(
"Incompatible dataframe 'df' to fit surrogate model of CBO."
)
self._opt.tell(x, [np.negative(yi).tolist() for yi in y])
[docs] def fit_generative_model(self, df, q=0.90, n_iter_optimize=0, n_samples=100):
"""Learn the distribution of hyperparameters for the top-``(1-q)x100%`` configurations and sample from this distribution. It can be used for transfer learning. For multiobjective problems, this function computes the top-``(1-q)x100%`` configurations in terms of their ranking with respect to pareto efficiency: all points on the first non-dominated pareto front have rank 1 and in general, points on the k'th non-dominated front have rank k.
Example Usage:
>>> search = CBO(problem, evaluator)
>>> search.fit_surrogate("results.csv")
Args:
df (str|DataFrame): a dataframe or path to CSV from a previous search.
q (float, optional): the quantile defined the set of top configurations used to bias the search. Defaults to ``0.90`` which select the top-10% configurations from ``df``.
n_iter_optimize (int, optional): the number of iterations used to optimize the generative model which samples the data for the search. Defaults to ``0`` with no optimization for the generative model.
n_samples (int, optional): the number of samples used to score the generative model.
Returns:
tuple: ``score, model`` which are a metric which measures the quality of the learned generated-model and the generative model respectively.
"""
# to make sdv optional
try:
import sdv
except ModuleNotFoundError:
raise deephyper.core.exceptions.MissingRequirementError(
"Installing 'sdv' is required to use 'fit_generative_model' please run 'pip install \"deephyper[sdv]\"'"
)
if type(df) is str and df[-4:] == ".csv":
df = pd.read_csv(df)
assert isinstance(df, pd.DataFrame)
if len(df) < 10:
raise ValueError(
f"The passed DataFrame contains only {len(df)} results when a minimum of 10 is required!"
)
# !avoid error linked to `n_components=10` a parameter of generative model used
q_max = 1 - 10 / len(df)
if q_max < q:
warnings.warn(
f"The value of q={q} is replaced by q_max={q_max} because a minimum of 10 results are required to perform transfer-learning!",
category=UserWarning,
)
q = q_max
# check single or multiple objectives
hp_cols = [k for k in df.columns if "p:" == k[:2]]
if "objective" in df.columns:
# filter failures
if pd.api.types.is_string_dtype(df.objective):
df = df[~df.objective.str.startswith("F")]
df.objective = df.objective.astype(float)
q_val = np.quantile(df.objective.values, q)
req_df = df.loc[df["objective"] > q_val]
else:
# filter failures
objcol = list(df.filter(regex=r"^objective_\d+$").columns)
for col in objcol:
if pd.api.types.is_string_dtype(df[col]):
df = df[~df[col].str.startswith("F")]
df[col] = df[col].astype(float)
top = non_dominated_set_ranked(-np.asarray(df[objcol]), 1.0 - q)
req_df = df.loc[top]
req_df = req_df[hp_cols]
req_df = req_df.rename(columns={k: k[2:] for k in hp_cols})
# constraints
scalar_constraints = []
for hp_name in self._problem.space:
if hp_name in req_df.columns:
hp = self._problem.space.get_hyperparameter(hp_name)
# TODO: Categorical and Ordinal are both considered non-ordered for SDV
# TODO: it could be useful to use the "category" type of Pandas and the ordered=True/False argument
# TODO: to extend the capability of SDV
if isinstance(hp, csh.CategoricalHyperparameter) or isinstance(
hp, csh.OrdinalHyperparameter
):
req_df[hp_name] = req_df[hp_name].astype("O")
else:
scalar_constraints.append(
sdv.constraints.ScalarRange(
hp_name, hp.lower, hp.upper, strict_boundaries=True
)
)
with warnings.catch_warnings():
warnings.simplefilter("ignore")
model = sdv.tabular.TVAE(constraints=scalar_constraints)
model.fit(req_df)
synthetic_data = model.sample(n_samples)
score = sdv.evaluation.evaluate(synthetic_data, req_df)
if n_iter_optimize > 0:
space = [
deephyper.skopt.space.Integer(1, 20, name="epochs"),
# deephyper.skopt.space.Integer(1, np.floor(req_df.shape[0]/10), name='batch_size'),
deephyper.skopt.space.Integer(1, 8, name="embedding_dim"),
deephyper.skopt.space.Integer(1, 8, name="compress_dims"),
deephyper.skopt.space.Integer(1, 8, name="decompress_dims"),
deephyper.skopt.space.Real(
10**-8, 10**-4, "log-uniform", name="l2scale"
),
deephyper.skopt.space.Integer(1, 5, name="loss_factor"),
]
def model_fit(params):
params["epochs"] = 10 * params["epochs"]
# params['batch_size'] = 10*params['batch_size']
params["embedding_dim"] = 2 ** params["embedding_dim"]
params["compress_dims"] = [
2 ** params["compress_dims"],
2 ** params["compress_dims"],
]
params["decompress_dims"] = [
2 ** params["decompress_dims"],
2 ** params["decompress_dims"],
]
model = sdv.tabular.TVAE(**params)
model.fit(req_df)
synthetic_data = model.sample(n_samples)
score = sdv.evaluation.evaluate(synthetic_data, req_df)
return -score, model
@use_named_args(space)
def objective(**params):
score, _ = model_fit(params)
return score
# run sequential optimization of generative model hyperparameters
opt = deephyper.skopt.Optimizer(space)
for i in range(n_iter_optimize):
x = opt.ask()
y = objective(x)
opt.tell(x, y)
logging.info(f"iteration {i}: {x} -> {y}")
min_index = np.argmin(opt.yi)
best_params = opt.Xi[min_index]
logging.info(
f"Min-Score of the SDV generative model: {opt.yi[min_index]}"
)
best_params = {d.name: v for d, v in zip(space, best_params)}
logging.info(
f"Best configuration for SDV generative model: {best_params}"
)
score, model = model_fit(best_params)
# we pass the learned generative model from sdv to the
# skopt Optimizer
self._opt_kwargs["model_sdv"] = model
return score, model
[docs] def fit_search_space(self, df, fac_numerical=0.125, fac_categorical=10):
"""Apply prior-guided transfer learning based on a DataFrame of results.
Example Usage:
>>> search = CBO(problem, evaluator)
>>> search.fit_surrogate("results.csv")
Args:
df (str|DataFrame): a checkpoint from a previous search.
fac_numerical (float): the factor used to compute the sigma of a truncated normal distribution based on ``sigma = max(1.0, (upper - lower) * fac_numerical)``. A small large factor increase exploration while a small factor increase exploitation around the best-configuration from the ``df`` parameter.
fac_categorical (float): the weight given to a categorical feature part of the best configuration. A large weight ``> 1`` increase exploitation while a small factor close to ``1`` increase exploration.
"""
if type(df) is str and df[-4:] == ".csv":
df = pd.read_csv(df)
assert isinstance(df, pd.DataFrame)
# check single or multiple objectives
if "objective" in df.columns:
# filter failures
if pd.api.types.is_string_dtype(df.objective):
df = df[~df.objective.str.startswith("F")]
df.objective = df.objective.astype(float)
else:
# filter failures
objcol = df.filter(regex=r"^objective_\d+$").columns
for col in objcol:
if pd.api.types.is_string_dtype(df[col]):
df = df[~df[col].str.startswith("F")]
df[col] = df[col].astype(float)
cst = self._problem.space
if type(cst) != CS.ConfigurationSpace:
logging.error(f"{type(cst)}: not supported for trainsfer learning")
res_df = df
res_df_names = res_df.columns.values
if "objective" in df.columns:
best_index = np.argmax(res_df["objective"].values)
best_param = res_df.iloc[best_index]
else:
best_index = non_dominated_set(
-np.asarray(res_df[objcol]), return_mask=False
)[0]
best_param = res_df.iloc[best_index]
cst_new = CS.ConfigurationSpace(seed=self._random_state.randint(0, 2**32))
hp_names = cst.get_hyperparameter_names()
for hp_name in hp_names:
hp = cst.get_hyperparameter(hp_name)
if hp_name in res_df_names:
if (
type(hp) is csh.UniformIntegerHyperparameter
or type(hp) is csh.UniformFloatHyperparameter
):
mu = best_param[hp.name]
lower = hp.lower
upper = hp.upper
sigma = max(1.0, (upper - lower) * fac_numerical)
if type(hp) is csh.UniformIntegerHyperparameter:
param_new = csh.NormalIntegerHyperparameter(
name=hp.name,
default_value=mu,
mu=mu,
sigma=sigma,
lower=lower,
upper=upper,
)
else: # type is csh.UniformFloatHyperparameter:
param_new = csh.NormalFloatHyperparameter(
name=hp.name,
default_value=mu,
mu=mu,
sigma=sigma,
lower=lower,
upper=upper,
)
cst_new.add_hyperparameter(param_new)
elif (
type(hp) is csh.CategoricalHyperparameter
or type(hp) is csh.OrdinalHyperparameter
):
if type(hp) is csh.OrdinalHyperparameter:
choices = hp.sequence
else:
choices = hp.choices
weights = len(choices) * [1.0]
index = choices.index(best_param[hp.name])
weights[index] = fac_categorical
norm_weights = [float(i) / sum(weights) for i in weights]
param_new = csh.CategoricalHyperparameter(
name=hp.name, choices=choices, weights=norm_weights
)
cst_new.add_hyperparameter(param_new)
else:
logging.warning(f"Not fitting {hp} because it is not supported!")
cst_new.add_hyperparameter(hp)
else:
logging.warning(
f"Not fitting {hp} because it was not found in the dataframe!"
)
cst_new.add_hyperparameter(hp)
# For conditions
for cond in cst.get_conditions():
if type(cond) == CS.AndConjunction or type(cond) == CS.OrConjunction:
cond_list = []
for comp in cond.components:
cond_list.append(self._return_cond(comp, cst_new))
if type(cond) is CS.AndConjunction:
cond_new = CS.AndConjunction(*cond_list)
elif type(cond) is CS.OrConjunction:
cond_new = CS.OrConjunction(*cond_list)
else:
logging.warning(f"Condition {type(cond)} is not implemented!")
else:
cond_new = self._return_cond(cond, cst_new)
cst_new.add_condition(cond_new)
# For forbiddens
for cond in cst.get_forbiddens():
if type(cond) is CS.ForbiddenAndConjunction:
cond_list = []
for comp in cond.components:
cond_list.append(self._return_forbid(comp, cst_new))
cond_new = CS.ForbiddenAndConjunction(*cond_list)
elif (
type(cond) is CS.ForbiddenEqualsClause
or type(cond) is CS.ForbiddenInClause
):
cond_new = self._return_forbid(cond, cst_new)
else:
logging.warning(f"Forbidden {type(cond)} is not implemented!")
cst_new.add_forbidden_clause(cond_new)
self._opt_kwargs["dimensions"] = cst_new
def _to_dict(self, x: list) -> dict:
"""Transform a list of hyperparameter values to a ``dict`` where keys are hyperparameters names and values are hyperparameters values.
Args:
x (list): a list of hyperparameter values.
Returns:
dict: a dictionnary of hyperparameter names and values.
"""
res = {}
hps_names = self._problem.hyperparameter_names
# to enforce native python types instead of numpy types
x = map(lambda xi: getattr(xi, "tolist", lambda: xi)(), x)
for hps_name, xi in zip(hps_names, x):
res[hps_name] = xi
return res
[docs]class AMBS(CBO):
"""AMBS is now deprecated and will be removed in the future use 'CBO' (Centralized Bayesian Optimization) instead!"""
def __init__(
self,
problem,
evaluator,
random_state: int = None,
log_dir: str = ".",
verbose: int = 0,
surrogate_model="RF",
acq_func: str = "UCB",
acq_optimizer: str = "auto",
kappa: float = 1.96,
xi: float = 0.001,
n_points: int = 10000,
filter_duplicated: bool = True,
update_prior: bool = False,
multi_point_strategy: str = "cl_max",
n_jobs: int = 1,
n_initial_points=10,
initial_points=None,
sync_communication: bool = False,
filter_failures: str = "mean",
**kwargs,
):
super().__init__(
problem,
evaluator,
random_state,
log_dir,
verbose,
surrogate_model,
acq_func,
acq_optimizer,
kappa,
xi,
n_points,
filter_duplicated,
update_prior,
multi_point_strategy,
n_jobs,
n_initial_points,
initial_points,
sync_communication,
filter_failures,
**kwargs,
)
import warnings
warnings.warn(
"'AMBS' is now deprecated and will be removed in the future use 'CBO' (Centralized Bayesian Optimization) instead!",
category=DeprecationWarning,
)
