π» My name is Marcel Robeer, and I am currently pursuing a PhD in Explainable Artificial Intelligence (XAI) at Utrecht University!
π€ My thesis projects and scientific research projects have resulted in several open-source Python packages:
- Explabox: {
Explore
|Examine
|Explain
|Expose
} your AI model with the explabox! - GlobalCausalAnalysis: Explaining Model Behavior with Global Causal Analysis (give a causal overview of how aspects such as task-related features, fairness and robustness relate to black-box model behavior) [xAI 2023 paper].
- text_explainability: A generic explainability architecture for explaining text machine learning models.
- text_sensitivity: Extension of text_explainability for sensitivity testing (robustness & fairness).
- CounterfactualGAN: Generating realistic natural language counterfactuals for classifiers and regressors, without requiring explainee intervention [EMNLP 2021 paper].
- ContrastiveExplanation: Contrastive and counterfactual explanations for machine learning with Foil Trees [WHI 2018 paper].
- VisualNarrator: Turns user stories into a conceptual model containing entities and relationships [RE 2016 paper].
π» Check out marcelrobeer.github.io for a full overview. See you there!