PROBE

While several approaches (e.g., see Implemented Counterfactual Methods) have been proposed to construct recourses for affected individuals, the recourses output by these methods either achieve low costs (i.e., ease-of-implementation) or robustness to small perturbations (i.e., noisy implementations of recourses), but not both due to the inherent trade-offs between the recourse costs and robustness. Our framework Probabilistically ROBust rEcourse (\texttt{PROBE}) lets users choose the probability with which a recourse could get invalidated (recourse invalidation rate) if small changes are made to the recourse i.e., the recourse is implemented somewhat noisily. To this end, we propose a novel objective function which simultaneously minimizes the gap between the achieved (resulting) and desired recourse invalidation rates, minimizes recourse costs, and also ensures that the resulting recourse achieves a positive model prediction.

Available Datasets

• Adult Data Set: Source
• COMPAS: Source
• Give Me Some Credit (GMC): Source

Implemented Counterfactual Methods

• Actionable Recourse (AR): Paper
• Diverse Counterfactual Explanations (DiCE): Paper
• Growing Sphere (GS): Paper
• Wachter: Paper
• ROAR: Paper
• ARAR: Paper

Provided Machine Learning Models

• ANN: Artificial Neural Network with 2 hidden layers and ReLU activation function
• LR: Linear Model with no hidden layer and no activation function

Dependence on CARLA

CARLA is a python library to benchmark counterfactual explanation and recourse models. It comes out-of-the box with commonly used datasets and various machine learning models. Documentation here and the corresponding NeurIPS paper here can be found using the corresponding links.

Installation

Using python directly or within activated virtual environment:

pip install -U pip setuptools wheel
pip install -e .

Requirements

• python3.7
• pip

Experiments

• To run our experiments, navigate to carla/recourse_invalidation_results/experiment/ and run recourseInvalidationRate.py
• To recreate some of the plots in the paper, run the following notebooks:
  • Bounds_Linear.ipynb, and
  • Bounds_ANN_Approx.ipynb.

Credit

This project was recently accepted to ICLR 2023. If you find our content helpful for your research, please cite:

@inproceedings{pawelczyk2023probabilistic,
      title={Probabilistically Robust Recourse: Navigating the Trade-offs between Costs and Robustness in Algorithmic Recourse},
      author={Martin Pawelczyk and Teresa Datta and Johannes van-den-Heuvel and Gjergji Kasneci and Himabindu Lakkaraju},
      booktitle={11th International Conference on Learning Representations (ICLR)},
      year={2023}
}