Location via proxy:   [ UP ]  
[Report a bug]   [Manage cookies]                
Skip to main content
Springer Nature Link
Log in

Good Counterfactuals and Where to Find Them: A Case-Based Technique for Generating Counterfactuals for Explainable AI (XAI)

  • Conference paper
  • First Online:
Case-Based Reasoning Research and Development (ICCBR 2020)

Part of the book series: Lecture Notes in Computer Science ((LNAI,volume 12311))

Included in the following conference series:

Abstract

Recently, a groundswell of research has identified the use of counterfactual explanations as a potentially significant solution to the Explainable AI (XAI) problem. It is argued that (i) technically, these counterfactual cases can be generated by permuting problem-features until a class-change is found, (ii) psychologically, they are much more causally informative than factual explanations, (iii) legally, they are GDPR-compliant. However, there are issues around the finding of “good” counterfactuals using current techniques (e.g. sparsity and plausibility). We show that many commonly-used datasets appear to have few “good” counterfactuals for explanation purposes. So, we propose a new case-based approach for generating counterfactuals, using novel ideas about the counterfactual potential and explanatory coverage of a case-base. The new technique reuses patterns of good counterfactuals, present in a case-base, to generate analogous counterfactuals that can explain new problems and their solutions. Several experiments show how this technique can improve the counterfactual potential and explanatory coverage of case-bases that were previously found wanting.

This is a preview of subscription content, log in via an institution to check access.

Access this chapter

Log in via an institution

Subscribe and save

Springer+ Basic
$34.99 /Month
  • Get 10 units per month
  • Download Article/Chapter or eBook
  • 1 Unit = 1 Article or 1 Chapter
  • Cancel anytime
Subscribe now

Buy Now

Chapter
USD 29.95
Price excludes VAT (USA)
  • Available as PDF
  • Read on any device
  • Instant download
  • Own it forever
eBook
USD 39.99
Price excludes VAT (USA)
  • Available as EPUB and PDF
  • Read on any device
  • Instant download
  • Own it forever
Softcover Book
USD 54.99
Price excludes VAT (USA)
  • Compact, lightweight edition
  • Dispatched in 3 to 5 business days
  • Free shipping worldwide - see info

Tax calculation will be finalised at checkout

Purchases are for personal use only

Institutional subscriptions

Similar content being viewed by others

Notes

  1. 1.

    This context assumes an existing (albeit opaque) model to which cases can be presented to find predictions/labels; all counterfactual-generation techniques make this assumption, though there is some discussion around whether the training data would also always be accessible (obviously, we assume the training-data/case-base is available).

  2. 2.

    Though NUNs have been studied in CBR (e.g., [22, 23]), few consider counterfactual cases (aka NUNs) for explanation; [24, 25] are exceptions but they viewed NUNs as being more important as confidence indicators with respect to decision boundaries.

  3. 3.

    Rare recent attempts include Laugel et al.’s [44] method to “justify” generated counterfactuals using nearest neighbors in the training data, and [29] finding “feasible paths” to counterfactuals in the dataset; both methods attempt to ground counterfactuals in prior experience.

  4. 4.

    We extensively tested this Blood Alcohol Content (BAC) case-base [24, 25], but cannot report it for reasons of space. Using a mechanical model for estimating BAC, we generated several master-case-bases from which we sampled 50+ specific case-bases; across all of these case-bases, to our astonishment, we repeatedly found the same absence of good counterfactuals.

  5. 5.

    More generally, for multi-class datasets, this adaptation can be modified to iterate over all ordered nearest neighbours with a different class to q, not just those with the same class as y′. This provides a larger pool of difference-feature values and increase the likelihood of locating a good counterfactual for q.

References

  1. Gunning, D.: Explainable artificial intelligence (XAI). Defense Advanced Research Projects Agency (DARPA), Web, vol. 2 (2017)

    Google Scholar 

  2. Gunning, D., Aha, D.W.: DARPA’s explainable artificial intelligence program. AI Mag. 40(2), 44–58 (2019)

    Article  Google Scholar 

  3. Goodman, B., Flaxman, S.: European Union regulations on algorithmic decision-making and a “right to explanation”. AI Mag. 38(3), 50–57 (2017)

    Article  Google Scholar 

  4. Wachter, S., Mittelstadt, B., Floridi, L.: Why a right to explanation of automated decision-making does not exist in the general data protection regulation. Int. Data Priv. Law 7(2), 76–99 (2017)

    Article  Google Scholar 

  5. Adadi, A., Berrada, M.: Peeking inside the black-box: a survey on Explainable Artificial Intelligence (XAI). IEEE Access 6, 52138–52160 (2018)

    Article  Google Scholar 

  6. Guidotti, R., Monreale, A., Ruggieri, S., Turini, F., Giannotti, F., Pedreschi, D.: A survey of methods for explaining black box models. ACM Comput. Surv. 51(5), 93 (2018)

    Google Scholar 

  7. Miller, T.: Explanation in artificial intelligence. Artif. Intell. 267, 1–38 (2019)

    Article  MATH  Google Scholar 

  8. Leake, D.B.: CBR in context: the present and future. In: Case-Based Reasoning: Experiences, Lessons, and Future Directions, pp. 3–30 (1996)

    Google Scholar 

  9. Leake, D., McSherry, D.: Introduction to the special issue on explanation in case-based reasoning. Artif. Intell. Rev. 24(2), 103–108 (2005)

    Article  Google Scholar 

  10. Sørmo, F., Cassens, J., Aamodt, A.: Explanation in case-based reasoning–perspectives and goals. Artif. Intell. Rev. 24(2), 109–143 (2005)

    Article  MATH  Google Scholar 

  11. Schoenborn, J.M., Althoff, K.D.: Recent trends in XAI: In: Case-Based Reasoning for the Explanation of intelligent systems (XCBR) Workshop (2019)

    Google Scholar 

  12. Lipton, Z.C.: The Mythos of model interpretability. Queue 16(3), 30 (2018)

    Article  Google Scholar 

  13. Kenny, E.M., Keane, M.T.: Twin-systems to explain neural networks using case-based reasoning. In: Proceedings of the 28th International Joint Conference on Artificial Intelligence, IJCAI 2019, pp. 326–333 (2019)

    Google Scholar 

  14. Keane, M.T., Kenny, E.M.: How case-based reasoning explains neural networks: a theoretical analysis of XAI using Post-Hoc explanation-by-example from a survey of ANN-CBR twin-systems. In: Bach, K., Marling, C. (eds.) ICCBR 2019. LNCS (LNAI), vol. 11680, pp. 155–171. Springer, Cham (2019). https://doi.org/10.1007/978-3-030-29249-2_11

    Chapter  Google Scholar 

  15. Byrne, R.M.J.: The Rational Imagination. MIT Press, Cambridge (2007)

    Google Scholar 

  16. Byrne, R.M.J.: Counterfactuals in explainable artificial intelligence (XAI): evidence from human reasoning. In: Proceedings of the Twenty-Eighth International Joint Conference on Artificial Intelligence, IJCAI 2019, pp. 6276–6282 (2019)

    Google Scholar 

  17. Wachter, S., Mittelstadt, B., Russell, C.: Counterfactual explanations without opening the black box: automated decisions and the GDPR. Harv. J. Law Tech. 31, 841 (2018)

    Google Scholar 

  18. Guidotti, R., Monreale, A., Giannotti, F., Pedreschi, D., Ruggieri S., Turini. F.: Factual and counterfactual explanations for black box decision making. IEEE Intell. Syst. 34(6), 14–23 (2019)

    Google Scholar 

  19. Smyth, B., Keane, M.T.: Remembering to forget. In: Proceedings of the 14th International Joint Conference on Artificial Intelligence, IJCAI 1995, pp. 377–382 (1995)

    Google Scholar 

  20. Smyth, B., McKenna, E.: Modelling the competence of case-bases. In: Smyth, B., Cunningham, P. (eds.) EWCBR 1998. LNCS, vol. 1488, pp. 208–220. Springer, Heidelberg (1998). https://doi.org/10.1007/BFb0056334

    Chapter  Google Scholar 

  21. Juarez, J.M., Craw, S., Lopez-Delgado, J.R., Campos, M.: Maintenance of case-bases: current algorithms after fifty years. In: Proceedings of the Twenty-Seventh International Joint Conference on Artificial Intelligence, IJCAI 2018, pp. 5457–5463 (2018)

    Google Scholar 

  22. Delany, S.J., Cunningham, P., Doyle, D., Zamolotskikh, A.: Generating estimates of classification confidence for a case-based spam filter. In: Muñoz-Ávila, H., Ricci, F. (eds.) ICCBR 2005. LNCS (LNAI), vol. 3620, pp. 177–190. Springer, Heidelberg (2005). https://doi.org/10.1007/11536406_16

    Chapter  Google Scholar 

  23. Kumar, R.R., Viswanath, P., Bindu, C.S.: Nearest neighbor classifiers: a review. Int. J. Comput. Intell. Res. 13(2), 303–311 (2017)

    Google Scholar 

  24. Cunningham, P., Doyle, D., Loughrey, J.: An evaluation of the usefulness of case-based explanation. In: Ashley, K.D., Bridge, D.G. (eds.) ICCBR 2003. LNCS (LNAI), vol. 2689, pp. 122–130. Springer, Heidelberg (2003). https://doi.org/10.1007/3-540-45006-8_12

    Chapter  MATH  Google Scholar 

  25. Nugent, C., Cunningham, P.: A case-based explanation system for black-box systems. Artif. Intell. Rev. 24(2), 163–178 (2005)

    Article  MATH  Google Scholar 

  26. Mittelstadt, B., Russell, C., Wachter, S.: Explaining explanations in AI. In: Proceedings of Conference on Fairness, Accountability, and Transparency, FAT 2019 (2019)

    Google Scholar 

  27. Pearl, J.: Causality, Cambridge University Press, Cambridge (2000)

    Google Scholar 

  28. Sokol, K., Flach, P.: Desiderata for interpretability: explaining decision tree predictions with counterfactuals. In: AAAI 20119, Doctoral Consortium, pp. 10035–10036 (2019)

    Google Scholar 

  29. Poyiadzi, R., Sokol, K., Santos-Rodriguez, R., De Bie, T., Flach, P.: FACE: feasible and actionable counterfactual explanations. In: Proceedings of the AAAI/ACM Conference on AI, Ethics, and Society, pp. 344–350 (2020). https://doi.org/10.1145/3375627.3375850

  30. Woodward, J.: Making Things Happen. Oxford University Press, Oxford (2003)

    Google Scholar 

  31. Van Fraassen, B.C.: The Scientific Image. Oxford University Press, Oxford (1980)

    Google Scholar 

  32. Kahneman, D., Miller, D.T.: Norm theory: comparing reality to its alternatives. Psychol. Rev. 93(2), 136–153 (1986)

    Article  Google Scholar 

  33. Mueller, S.T., Hoffman, R.R., Clancey, W.J., Emery, A.K., Klein, G.: Explanation in human-AI systems. Florida Institute for Human and Machine Cognition (2019)

    Google Scholar 

  34. Dodge, J., Liao, Q.V., Zhang, Y., Bellamy, R.K., Dugan, C.: Explaining models: an empirical study of how explanations impact fairness judgment. In: Proceedings of the 24th International Conference on Intelligent User Interfaces, pp. 275–285 (2019)

    Google Scholar 

  35. Miller, T.: Contrastive explanation. arXiv preprint arXiv:1811.03163 (2018)

  36. Russell, C., Kusner, M.J., Loftus, J., Silva, R.: When worlds collide: integrating different counterfactual assumptions in fairness. In: Advances in Neural Information Processing Systems, pp. 6414–6423 (2017)

    Google Scholar 

  37. Ribeiro, M.T., Singh, S., Guestrin, C.: Why should I trust you?. In: Proceedings of the 22nd ACM SIGKDD, pp. 1135–1144. ACM (2016)

    Google Scholar 

  38. Pedreschi, D., Giannotti, F., Guidotti, R., Monreale, A., Ruggieri, S., Turini, F.: Meaningful explanations of Black Box AI decision systems. In: Proceedings of AAAI 2019 (2019)

    Google Scholar 

  39. Mothilal, R.K., Sharma, A., Tan, C.: Explaining machine learning classifiers through diverse counterfactual explanations. In: Proceedings of the 2020 Conference on Fairness, Accountability, and Transparency, FAT 2020, pp. 607–617 (2020)

    Google Scholar 

  40. McGrath, R., et al.: Interpretable credit application predictions with counterfactual explanations. In: NIP Workshop on Challenges and Opportunities for AI in Financial Services, Montreal, Canada (2018)

    Google Scholar 

  41. Miller, G.A.: The magical number seven, plus or minus two: some limits on our capacity for processing information. Psychol. Rev. 63(2), 81 (1956)

    Article  Google Scholar 

  42. Alvarez, G., Cavanagh, P.: The capacity of visual STM is set both by visual information load and by number of objects. Psychol. Sci. 15, 106–111 (2004)

    Article  Google Scholar 

  43. Medin, D.L., Wattenmaker, W.D., Hampson, S.E.: Family resemblance, conceptual cohesiveness, and category construction. Cogn. Psychol. 19(2), 242–279 (1987)

    Article  Google Scholar 

  44. Laugel, T., Lesot, M.J., Marsala, C., Renard, X., Detyniecki, M.: The dangers of post-hoc interpretability: unjustified counterfactual explanations. In: Proceedings of the 28th International Joint Conference on Artificial Intelligence, IJCAI 2019, pp. 2801–2807 (2019)

    Google Scholar 

  45. Dua, D., Graff, C.: UCI Machine Learning Repository University of California, School of Information and Computer Science, Irvine, CA. http://archive.ics.uci.edu/ml (2019)

  46. Lieber, J., Nauer, E., Prade, H.: Improving analogical extrapolation using case pair competence. In: Bach, K., Marling, C. (eds.) ICCBR 2019. LNCS (LNAI), vol. 11680, pp. 251–265. Springer, Cham (2019). https://doi.org/10.1007/978-3-030-29249-2_17

    Chapter  Google Scholar 

  47. Veale, T., Keane, M.T.: The competence of sub-optimal theories of structure mapping on hard analogies. In: International Joint Conference on Artificial Intelligence, pp. 232–237 (1997)

    Google Scholar 

  48. Keane, M.T.: Analogical asides on case-based reasoning. In: Wess, S., Althoff, K.D., Richter, M.M. (eds.) EWCBR 1993. LNCS, vol. 837, pp. 21–32. Springer, Heidelberg (1994). https://doi.org/10.1007/3-540-58330-0_74

    Chapter  Google Scholar 

  49. Karimi, A.H., Barthe, G., Balle, B., Valera, I.: Model-agnostic counterfactual explanations for consequential decisions. In: Proceedings of the 23rd International Conference on Artificial Intelligence and Statistics, AISTATS 2020, Palermo, Italy, vol. 108. PMLR (2020)

    Google Scholar 

  50. Sokol, K., Flach, P.: Explainability fact sheets: a framework for systematic assessment of explainable approaches. In: Proceedings of the 2020 Conference on Fairness, Accountability, and Transparency, FAT 2020, pp. 56–67 (2020)

    Google Scholar 

Download references

Acknowledgements

This paper emanated from research funded by (i) Science Foundation Ireland (SFI) to the Insight Centre for Data Analytics under Grant Number 12/RC/2289_P2 and (ii) SFI and the Department of Agriculture, Food and Marine on behalf of the Government of Ireland to the VistaMilk SFI Research Centre under Grant Number 16/RC/3835.

Author information

Authors and Affiliations

  1. School of Computer Science, University College Dublin, Dublin, Ireland

    Mark T. Keane & Barry Smyth

  2. Insight Centre for Data Analytics, University College Dublin, Dublin, Ireland

    Mark T. Keane & Barry Smyth

  3. VistaMilk SFI Research Centre, University College Dublin, Dublin, Ireland

    Mark T. Keane

Authors
  1. Mark T. Keane
    View author publications

    You can also search for this author in PubMed Google Scholar

  2. Barry Smyth
    View author publications

    You can also search for this author in PubMed Google Scholar

Corresponding author

Correspondence to Barry Smyth .

Editor information

Editors and Affiliations

  1. School of Computer Science, University of Auckland, Auckland, New Zealand

    Ian Watson

  2. College of Computing and Informatics, Drexel University, Philadelphia, PA, USA

    Rosina Weber

Rights and permissions

Reprints and permissions

Copyright information

© 2020 Springer Nature Switzerland AG

About this paper

Check for updates. Verify currency and authenticity via CrossMark

Cite this paper

Keane, M.T., Smyth, B. (2020). Good Counterfactuals and Where to Find Them: A Case-Based Technique for Generating Counterfactuals for Explainable AI (XAI). In: Watson, I., Weber, R. (eds) Case-Based Reasoning Research and Development. ICCBR 2020. Lecture Notes in Computer Science(), vol 12311. Springer, Cham. https://doi.org/10.1007/978-3-030-58342-2_11

Download citation

  • DOI: https://doi.org/10.1007/978-3-030-58342-2_11

  • Published:

  • Publisher Name: Springer, Cham

  • Print ISBN: 978-3-030-58341-5

  • Online ISBN: 978-3-030-58342-2

  • eBook Packages: Computer ScienceComputer Science (R0)

Publish with us

Policies and ethics

Search

Navigation