pertcf-explainer
This repository provides reproducible benchmarking experiments of 'PertCF: A Perturbation-Based Counterfactual Generation Approach' paper and open-source implementation of the proposed PertCF method.
Science Score: 44.0%
This score indicates how likely this project is to be science-related based on various indicators:
-
✓CITATION.cff file
Found CITATION.cff file -
✓codemeta.json file
Found codemeta.json file -
✓.zenodo.json file
Found .zenodo.json file -
○DOI references
-
○Academic publication links
-
○Committers with academic emails
-
○Institutional organization owner
-
○JOSS paper metadata
-
○Scientific vocabulary similarity
Low similarity (15.5%) to scientific vocabulary
Repository
This repository provides reproducible benchmarking experiments of 'PertCF: A Perturbation-Based Counterfactual Generation Approach' paper and open-source implementation of the proposed PertCF method.
Basic Info
Statistics
- Stars: 1
- Watchers: 1
- Forks: 0
- Open Issues: 0
- Releases: 0
Metadata Files
README.md
PertCF-Explainer
This repository provides reproducible benchmarking experiments of 'PertCF: A Perturbation-Based Counterfactual Generation Approach' paper and open-source implementation of the proposed PertCF method.
PertCF is an iterative perturbation-based counterfactual generation method that combines the strengths of perturbation-based counterfactual generation and feature attribution methods to generate high-quality, stable, and interpretable counterfactual explanations.
Instructions to run the contribution
Step 1: Clone the repository
Step 2: Build myCBR API
This project requires to run the myCBR Rest API. Therefore you will need to have Java 8 JRE on your computer and download the following jar: https://folk.idi.ntnu.no/kerstinb/mycbr/mycbr-rest/mycbr-rest-2.0.jar Make sure to add the jar file to the myCBR folder.
This jar is built using this repository https://github.com/ntnu-ai-lab/mycbr-rest
Step 3: Starting the API
We assume that you have cloned the GitHub repository and added the mycbr-rest-2.0.jar into the mycbr folder of this repository. Then in the terminal, navigate to the cloned project folder.
Then go into the mycbr folder and run the following command:
java -DMYCBR.PROJECT.FILE=PROJECT_FILE.prj -jar mycbr-rest-2.0.jar
This command will start a local webserver, deploy the myCBR project and expose it's endpoints via a Rest API. We assume that this server is running in the background while the experiments notebooks are run.
Step 4: Accesing API
Once the installation is done and the API is running, the API will be accessibleat http://localhost:8080/swagger-ui.html#. To verify that the API is up and running, please check that you can see the swagger documentation page for the API.
Step 5: Run the experiments
Necessary files to run the experiments
To run the experiments we provide 'SouhtGermanCredit' and 'UserKnowledgeModeling' folders with a number of files under the 'Experiments' folder. The folders consist of experiments and the following folders to reproduce the experiments:
Data/: contains .csv file/s needed for the experiments.
Model/: contains the .pkl files for the BB models used in the experiments.
Output/: contains result files of the experiments.
temp_files/: contains all other files that are generated and/or used during experimenting.
'SouhtGermanCredit' and 'UserKnowledgeModeling' folders also include reproducible and adaptable ordered set of experiments as ipython notebooks.
This repository also includes followings:
myCBR_projects/: contains myCBR project files that are modelled for the experiments.
mycbrpyapi.py: contains myCBR-rest API connection and calls
PertCF.py: contains Explainer Class that implements PertCF including experimental setup
Owner
- Name: Betül Bayrak
- Login: b-bayrak
- Kind: user
- Location: Norway
- Repositories: 1
- Profile: https://github.com/b-bayrak
Citation (CITATION.cff)
cff-version: 1.2.0
message: "Please cite it as below."
authors:
- family-names: "Bayrak"
given-names: "Betül"
orcid: "https://orcid.org/0000-0002-0554-9823"
affiliation: "Norwegian University of Science and Technology"
- family-names: "Bach"
given-names: "Kerstin"
orcid: "https://orcid.org/0000-0002-4256-7676"
affiliation: "Norwegian University of Science and Technology"
title: "PertCF-Explainer"
version: 1.5.1
date-released: 2023-09-11
url: "https://github.com/b-bayrak/PertCF-Explainer"
preferred-citation:
authors:
- family-names: "Bayrak"
given-names: "Betül"
orcid: "https://orcid.org/0000-0002-0554-9823"
affiliation: "Norwegian University of Science and Technology"
- family-names: "Bach"
given-names: "Kerstin"
orcid: "https://orcid.org/0000-0002-4256-7676"
affiliation: "Norwegian University of Science and Technology"
title: "PertCF: A Perturbation-Based Counterfactual Generation Approach"
year: 2023
conference:
name: "International Conference on Innovative Techniques and Applications of Artificial Intelligence"
city: "Cambridge"
country: "UK"
pages: 174--187
organization: "Springer"
GitHub Events
Total
Last Year
Issues and Pull Requests
Last synced: 11 months ago
All Time
- Total issues: 0
- Total pull requests: 0
- Average time to close issues: N/A
- Average time to close pull requests: N/A
- Total issue authors: 0
- Total pull request authors: 0
- Average comments per issue: 0
- Average comments per pull request: 0
- Merged pull requests: 0
- Bot issues: 0
- Bot pull requests: 0
Past Year
- Issues: 0
- Pull requests: 0
- Average time to close issues: N/A
- Average time to close pull requests: N/A
- Issue authors: 0
- Pull request authors: 0
- Average comments per issue: 0
- Average comments per pull request: 0
- Merged pull requests: 0
- Bot issues: 0
- Bot pull requests: 0