Difan Deng

Rolling Back PR #773 when ConfigSpace's issue is fixed

3

We need to roll back the merged PR https://github.com/automl/SMAC3/pull/773 once the bug at https://github.com/automl/ConfigSpace/issues/194 is fixed

intial designs raise an error if the generated vector conflicts the forbidden clauses

5

#### Description If forbidden clauses exist in smac's configspace and initial design generators generates a vector that conflicts the forbidden clauses, rather then ignore and skip that, SMAC will raise...

we could build an example making use of [this package](https://github.com/RobertTLange/mle-hyperopt) to make the interface cleaner and generate nice outputs.

NOTE: ISSUES ARE NOT FOR CODE HELP - Ask for Help at https://stackoverflow.com Your issue may already be reported! Also, please search on the [issue tracker](../) before creating one. *...

Nested Conditions result in incorrect deactivation

4

```python from ConfigSpace.configuration_space import ConfigurationSpace, Configuration from ConfigSpace.hyperparameters import CategoricalHyperparameter from ConfigSpace.conditions import EqualsCondition, AndConjunction, OrConjunction from ConfigSpace.util import get_one_exchange_neighbourhood import numpy as np rng = np.random.RandomState(1) cs = ConfigurationSpace()...

When combining Integral hyperparameter and conditions, sometimes it might result in an invalid configuration: ``` from ConfigSpace import Configuration, ConfigurationSpace, GreaterThanCondition from ConfigSpace import UniformIntegerHyperparameter import ConfigSpace num_layer = UniformIntegerHyperparameter('num_layer',...

[Time series Forecasting] Continuous Ranked Probablity Score (CRPS) loss for probablity network ouput type

2

We could consider implementing CPRS an alternative training loss for the probabilistic network output type. INFO: https://www.lokad.com/continuous-ranked-probability-score

## Types of changes - [ ] Breaking change (fix or feature that would cause existing functionality to not work as expected) - [x] Bug fix (non-breaking change which fixes...

We need to create an example showing that each fold of cross-validation can be considered as an instance in SMAC, similar to https://github.com/automl/SMAC3/blob/main/examples/2_multi_fidelity/2_sgd_datasets.py

Some tests, e.g., tests/test_model/test_gp.py::test_gp_on_sklearn_data tests/test_model/test_gp_mcmc.py::test_predict_with_actual_values tests/test_model/test_gp_mcmc.py::test_gp_on_sklearn_data tests/test_model/test_rf.py::test_rf_on_sklearn_data We need to check that if they still work and replace the `load_boston` with `load_diabetes`