DiCE
DiCE copied to clipboard
interpretml
Generate Diverse Counterfactual Explanations for any machine learning model.
Installed dice-ml from pypi. ``` import dice_ml from dice_ml.utils import helpers # helper functions d = dice_ml.Data(dataframe=helpers.load_adult_income_dataset(), continuous_features=['age', 'hours_per_week'], outcome_name='income') m = dice_ml.Model(model_path=dice_ml.utils.helpers.get_adult_income_modelpath()) exp = dice_ml.Dice(d,m) ``` Getting the following...
DiCE seems awesome. Thank you for your work on it! I am trying to use DiCE with XGBoost/LightGBM but I am getting some unexpected behaviour. First and foremost, DiCE seems...
**[READY FOR MERGE]** This PR adds support for private data interface for sklearn ML models. dice-kdtree is not supported since it requires access to the training data to output the...
Hi, I have tried to use three methods to get CounterfactualExamples in Jupyter notebook. But I find the **_new_outcome_** is wrong sometimes when _**desired_class=="opposite".**_ And I check your code, I...
This is a modification to unify the type comparison method. Previously, there were two methods - `if type(ABC) is str:` - `if isinstance(ABC, str):` This fix is related to coding...
**[READY TO MERGE]** Refactors deep learning model and explainers so that they are compatible with the sklearn explainers. sklearn explainers do not need access to gradient, so they should also...
This is an extension of the Gradient-based method to handle the multi class case (for tensorflow2 only). The method is added as a separate class, _DiceTensorFlow2Multi_. It can be used...
Hi. Could you add another option or method to save counterfactuals? to_json() method doesn't work when all the variables are categorical and they are in float format. (given the message,...
the method compute_proximity_loss() from dice_genetic.py returns an error when the dataset does not contain continous features. ``` def compute_proximity_loss(self, x_hat_unnormalized, query_instance_normalized): """Compute weighted distance between two vectors.""" x_hat = self.data_interface.normalize_data(x_hat_unnormalized)...
Link mistake in VAE method: Gradient-based methods - An explicit loss-based method described in [Mothilal et al. (2020)](https://arxiv.org/abs/1905.07697) (Default for deep learning models). - A Variational AutoEncoder (VAE)-based method described...
Metadata
Owner
Metadata
Generate Diverse Counterfactual Explanations for any machine learning model.