Research Proposal

You may not like it, but this is what counterfactuals look like

• Show DiCE for weak MLP
• Show Latent for same weak MLP
• Latent can be manipulated:
  • train biased model
  • train VAE with biased variable removed/attacked (use Boston housing dataset)
  • hypothesis: will generate bias-free explanations

Introduction to Conformal Prediction

• distribution-free, model-agnostic and scalable approach to predictive uncertainty quantification

Post-hoc

• Take any fitted model and turn it into a conformal model using calibration data.

Intrinsic — Conformal Training [MAYBE]

• Model explicitly trained for conformal prediction.

Conformal Counterfactuals

• Realistic counterfactuals by minimizing predictive uncertainty (Schut et al. 2021).
• Problem: restricted to Bayesian models.
• Solution: post-hoc predictive uncertainty quantification.
• Conformal prediction is instance-based. So is CE.
• Does the coverage guarantee carry over to counterfactuals?

Research Questions

• Is CP alone enough to ensure realistic counterfactuals?
• Do counterfactuals improve further as the models get better?
• Do counterfactuals get more realistic as coverage
• What happens as we vary coverage and setsize?
• What happens as we improve the model robustness?
• What happens as we improve the model’s ability to incorporate predictive uncertainty (deep ensemble, laplace)?

Experiments

• Maybe: conformalised Laplace
• Benchmarking:
  • add PROBE into the mix
  • compare travel costs to domain shits.

References

Schut, Lisa, Oscar Key, Rory Mc Grath, Luca Costabello, Bogdan Sacaleanu, Yarin Gal, et al. 2021. “Generating Interpretable Counterfactual Explanations By Implicit Minimisation of Epistemic and Aleatoric Uncertainties.” In International Conference on Artificial Intelligence and Statistics, 1756–64. PMLR.