2020

Algorithmic Recourse: from Counterfactual Explanations to Interventions

Karimi, Amir-Hossein, Schölkopf, Bernhard, Valera, Isabel

Understand

As machine learning is increasingly used to inform consequential decision-making (e.g., pre-trial bail and loan approval), it becomes important to explain how the system arrived at its decision, and also suggest actions to achieve a favorable decision.

Counterfactual explanations -- "how the world would have (had) to be different for a desirable outcome to occur" -- aim to satisfy these criteria.
Existing works have primarily focused on designing algorithms to obtain counterfactual explanations for a wide range of settings.
However, one of the main objectives of "explanations as a means to help a data-subject act rather than merely understand" has been overlooked.

Built on

The comprehensibility manifesto
Yves Kodratoff. 1994 · 1994
Earlier work this paper cites.
A probabilistic calculus of actions
Judea Pearl. 1994 · 1994
Earlier work this paper cites.
Causality: models, reasoning and inference . Vol. 29
Judea Pearl. 2000 · 2000
Earlier work this paper cites.
Causation, prediction, and search
Peter Spirtes, Clark N Glymour, Richard Scheines, and David Heckerman. 2000 · 2000
Earlier work this paper cites.
Varieties of causal intervention. In Pacific Rim International Conference on Artificial Intelligence . Springer, 322–331
Kevin B Korb, Lucas R Hope, Ann E Nicholson, and Karl Axnick. 2004 · 2004
Earlier work this paper cites.
Learning interpretable models
Stefan Rüping. 2006 · 2006
Earlier work this paper cites.
Causation and intervention
Frederick Eberhardt. 2007 · 2007
Earlier work this paper cites.
Interventions and causal inference
Frederick Eberhardt and Richard Scheines. 2007 · 2007
Earlier work this paper cites.
UCI machine learning repository
Kevin Bache and Moshe Lichman. 2013 · 2013
Earlier work this paper cites.
Causal responsibility and counterfactuals
David A Lagnado, Tobias Gerstenberg, and Ro’i Zultan. 2013 · 2013
Earlier work this paper cites.
Structural counterfactuals: A brief introduction
Judea Pearl. 2013 · 2013
Earlier work this paper cites.
Indicative Conditionals
Dorothy Edgington. 2014 · 2014
Earlier work this paper cites.
Causal inference in statistics: A primer
Judea Pearl, Madelyn Glymour, and Nicholas P Jewell. 2016 · 2016
Earlier work this paper cites.
Interventions over predictions: Reframing the ethical debate for actuarial risk assessment
Original
Chelsea Barabas, Karthik Dinakar, Joichi Ito, Madars Virza, and Jonathan Zittrain. 2017 · 2017
Earlier work this paper cites.
Towards a rigorous science of interpretable machine learning
Original
Finale Doshi-Velez and Been Kim. 2017 · 2017
Earlier work this paper cites.
Introduction to the foundations of causal discovery
Frederick Eberhardt. 2017 · 2017
Earlier work this paper cites.
Avoiding discrimination through causal reasoning. In Advances in Neural Information Processing Systems . 656–666
Niki Kilbertus, Mateo Rojas Carulla, Giambattista Parascandolo, Moritz Hardt, Dominik Janzing, and Bernhard Schölkopf. 2017 · 2017
Earlier work this paper cites.
Counterfactual fairness. In Advances in Neural Information Processing Systems . 4066–4076
Matt J Kusner, Joshua Loftus, Chris Russell, and Ricardo Silva. 2017 · 2017
Earlier work this paper cites.

Similar

Inverse Classification for Comparison-based Interpretability in Machine Learning
Original
Thibault Laugel, Marie-Jeanne Lesot, Christophe Marsala, Xavier Renard, and Marcin Detyniecki. 2017 · 2017
Cited alongside, same era.
Elements of causal inference
Jonas Peters, Dominik Janzing, and Bernhard Schölkopf. 2017 · 2017
Cited alongside, same era.
Causal consistency of structural equation models
Original
Paul K Rubenstein, Sebastian Weichwald, Stephan Bongers, Joris M Mooij, Dominik Janzing, Moritz Grosse-Wentrup, and Bernhard Schölkopf. 2017 · 2017
Cited alongside, same era.
When worlds collide: integrating different counterfactual assumptions in fairness. In Advances in Neural Information Processing Systems . 6414–6423
Chris Russell, Matt J Kusner, Joshua Loftus, and Ricardo Silva. 2017 · 2017
Cited alongside, same era.
Counterfactual explanations without opening the black box: Automated decisions and the GDPR
Sandra Wachter, Brent Mittelstadt, and Chris Russell. 2017 · 2017
Cited alongside, same era.
Explanations based on the missing: Towards contrastive explanations with pertinent negatives. In Advances in Neural Information Processing Systems . 592–603
Amit Dhurandhar, Pin-Yu Chen, Ronny Luss, Chun-Chen Tu, Paishun Ting, Karthikeyan Shanmugam, and Payel Das. 2018 · 2018
Cited alongside, same era.
Local rule-based explanations of black box decision systems
Original
Riccardo Guidotti, Anna Monreale, Salvatore Ruggieri, Dino Pedreschi, Franco Turini, and Fosca Giannotti. 2018 · 2018
Cited alongside, same era.
Eddie Murphy and the dangers of counterfactual causal thinking about detecting racial discrimination
Issa Kohler-Hausmann. 2018 · 2018
Cited alongside, same era.
The mythos of model interpretability
Zachary C Lipton. 2018 · 2018
Cited alongside, same era.
Causal discovery algorithms: A practical guide
Daniel Malinsky and David Danks. 2018 · 2018
Cited alongside, same era.
Abstracting causal models. In Proceedings of the AAAI Conference on Artificial Intelligence , Vol. 33. 2678–2685
Sander Beckers and Joseph Y Halpern. 2019 · 2019
Cited alongside, same era.
Path-specific counterfactual fairness. In Proceedings of the AAAI Conference on Artificial Intelligence , Vol. 33. 7801–7808
Silvia Chiappa. 2019 · 2019
Cited alongside, same era.

Then

Preserving Causal Constraints in Counterfactual Explanations for Machine Learning Classifiers
Original
Divyat Mahajan, Chenhao Tan, and Amit Sharma. 2019 · 2019
Later among the works it cites.
Explanation in artificial intelligence: Insights from the social sciences
Tim Miller. 2019 · 2019
Later among the works it cites.
DiCE: Explaining Machine Learning Classifiers through Diverse Counterfactual Explanations
Original
Ramaravind Kommiya Mothilal, Amit Sharma, and Chenhao Tan. 2019 · 2019
Later among the works it cites.
Definitions, methods, and applications in interpretable machine learning
W James Murdoch, Chandan Singh, Karl Kumbier, Reza Abbasi-Asl, and Bin Yu. 2019 · 2019
Later among the works it cites.
FACE: Feasible and Actionable Counterfactual Explanations
Original
Rafael Poyiadzi, Kacper Sokol, Raul Santos-Rodriguez, Tijl De Bie, and Peter Flach. 2019 · 2019
Later among the works it cites.
Synthesizing Action Sequences for Modifying Model Decisions
Original
Goutham Ramakrishnan, Yun Chan Lee, and Aws Albargouthi. 2019 · 2019
Later among the works it cites.
Stop explaining black box machine learning models for high stakes decisions and use interpretable models instead
Cynthia Rudin. 2019 · 2019
Later among the works it cites.
Efficient Search for Diverse Coherent Explanations. In Proceedings of the Conference on Fairness, Accountability, and Transparency (FAT* ’19) . ACM, 20–28
Chris Russell. 2019 · 2019
Later among the works it cites.
CERTIFAI: Counterfactual Explanations for Robustness, Transparency, Interpretability, and Fairness of Artificial Intelligence models
Original
Shubham Sharma, Jette Henderson, and Joydeep Ghosh. 2019 · 2019
Later among the works it cites.
Counterfactuals
William Starr. 2019a · 2019
Later among the works it cites.
Counterfactuals
William Starr. 2019b · 2019
Later among the works it cites.
Actionable recourse in linear classification. In Proceedings of the Conference on Fairness, Accountability, and Transparency . ACM, 10–19
Berk Ustun, Alexander Spangher, and Yang Liu. 2019 · 2019
Later among the works it cites.
The What-If Tool: Interactive Probing of Machine Learning Models
James Wexler, Mahima Pushkarna, Tolga Bolukbasi, Martin Wattenberg, Fernanda Viégas, and Jimbo Wilson. 2019 · 2019
Later among the works it cites.
The hidden assumptions behind counterfactual explanations and principal reasons. In Proceedings of the 2020 Conference on Fairness, Accountability, and Transparency . 80–89
Solon Barocas, Andrew D Selbst, and Manish Raghavan. 2020 · 2020
Closest in time.
Robustness in machine learning explanations: does it matter?. In Proceedings of the 2020 Conference on Fairness, Accountability, and Transparency . 640–647
Leif Hancox-Li. 2020 · 2020
Closest in time.
Model-agnostic counterfactual explanations for consequential decisions. In International Conference on Artificial Intelligence and Statistics . 895–905
Amir-Hossein Karimi, Gilles Barthe, Borja Balle, and Isabel Valera. 2020 · 2020
Closest in time.
The philosophical basis of algorithmic recourse. In Proceedings of the Conference on Fairness, Accountability, and Transparency . ACM
Suresh Venkatasubramanian and Mark Alfano. 2020 · 2020
Closest in time.

Beyond the bibliography

alphaXiv searches the wider corpus for related work and actual follow-ups.

Open on alphaXiv

alphaXiv is searching for related work…

Reading the bibliography…