Deep Influence Diagrams: An Interpretable and Robust Decision Support System

  • Hal James CooperEmail author
  • Garud Iyengar
  • Ching-Yung Lin
Conference paper
Part of the Lecture Notes in Business Information Processing book series (LNBIP, volume 353)


Interpretable decision making frameworks allow us to easily endow agents with specific goals, risk tolerances, and understanding. Existing decision making systems either forgo interpretability, or pay for it with severely reduced efficiency and large memory requirements. In this paper, we outline DeepID, a neural network approximation of Influence Diagrams, that avoids both pitfalls. We demonstrate how the framework allows for the introduction of robustness in a very transparent and interpretable manner, without increasing the complexity class of the decision problem.


Agent systems and collective intelligence Robust systems Interpretable systems 


  1. 1.
    Abad, C., Iyengar, G.: Portfolio selection with multiple spectral risk constraints. SIAM J. Financ. Math. 6(1), 467–486 (2015)MathSciNetCrossRefGoogle Scholar
  2. 2.
    Artzner, P., Delbaen, F., Eber, J.M., Heath, D.: Coherent measures of risk. Math. Financ. 9(3), 203–228 (1999)MathSciNetCrossRefGoogle Scholar
  3. 3.
    Bielza, C., Gomez, M., Shenoy, P.P.: A review of representation issues and modeling challenges with influence diagrams. Omega 39(3), 227–241 (2011)CrossRefGoogle Scholar
  4. 4.
    Boucherie, R.J., Van Dijk, N.M.: Markov Decision Processes in Practice, vol. 248. Springer, Cham (2017). Scholar
  5. 5.
    Cooper, H., Iyengar, G., Lin, C.Y.: Interpretable robust decision making. In: International Conference on Autonomous Agents and Multiagent Systems, Stockholm (2018)Google Scholar
  6. 6.
    Diez, F.J., et al.: Markov influence diagrams: a graphical tool for cost-effectiveness analysis. Med. Decis. Making 37(2), 183–195 (2017)CrossRefGoogle Scholar
  7. 7.
    Everitt, T., Ortega, P.A., Barnes, E., Legg, S.: Understanding agent incentives using causal influence diagrams, Part I: single action settings. arXiv preprint arXiv:1902.09980 (2019)
  8. 8.
    Gomez, M., Bielza, C., del Pozo, J.A.F., Rios-Insua, S.: A graphical decision-theoretic model for neonatal jaundice. Med. Decis. Making 27(3), 250–265 (2007)CrossRefGoogle Scholar
  9. 9.
    Goodfellow, I., Bengio, Y., Courville, A.: Deep Learning. MIT Press (2016).
  10. 10.
    Goodfellow, I., et al.: Generative adversarial nets. In: Advances in Neural Information Processing Systems, pp. 2672–2680 (2014)Google Scholar
  11. 11.
    Gulrajani, I., Ahmed, F., Arjovsky, M., Dumoulin, V., Courville, A.: Improved training of Wasserstein GANs. In: Proceedings of the 31st International Conference on Neural Information Processing Systems, pp. 5769–5779. Curran Associates Inc. (2017)Google Scholar
  12. 12.
    Howard, R.A.: Readings on the Principles and Applications of Decision Analysis, vol. 1. Strategic Decisions Group (1983)Google Scholar
  13. 13.
    Iyengar, G.N.: Robust dynamic programming. Math. Oper. Res. 30(2), 257–280 (2005)MathSciNetCrossRefGoogle Scholar
  14. 14.
    Jensen, F.V., Nielsen, T.D.: Probabilistic decision graphs for optimization under uncertainty. Ann. Oper. Res. 204(1), 223–248 (2013)MathSciNetCrossRefGoogle Scholar
  15. 15.
    Kingma, D.P., Welling, M.: Auto-encoding variational Bayes. arXiv preprint arXiv:1312.6114 (2013)
  16. 16.
    Larochelle, H., Murray, I.: The neural autoregressive distribution estimator. In: AISTATS, vol. 1, p. 2 (2011)Google Scholar
  17. 17.
    Maddison, C.J., Mnih, A., Teh, Y.W.: The concrete distribution: a continuous relaxation of discrete random variables. arXiv preprint arXiv:1611.00712 (2016)
  18. 18.
    Magni, P., Quaglini, S., Marchetti, M., Barosi, G.: Deciding when to intervene: a Markov decision process approach. Int. J. Med. Informatics 60(3), 237–253 (2000)CrossRefGoogle Scholar
  19. 19.
    Mirza, M., Osindero, S.: Conditional generative adversarial nets. arXiv preprint arXiv:1411.1784 (2014)
  20. 20.
    Nilim, A., El Ghaoui, L.: Robust control of Markov decision processes with uncertain transition matrices. Oper. Res. 53(5), 780–798 (2005)MathSciNetCrossRefGoogle Scholar
  21. 21.
    Virto, M.A., Martin, J., Insua, D.R., Moreno-Diaz, A.: Approximate solutions of complex influence diagrams through MCMC methods. In: Probabilistic Graphical Models (2002)Google Scholar

Copyright information

© Springer Nature Switzerland AG 2019

Authors and Affiliations

  • Hal James Cooper
    • 1
    Email author
  • Garud Iyengar
    • 1
  • Ching-Yung Lin
    • 1
    • 2
  1. 1.Columbia UniversityNew YorkUSA
  2. 2.Graphen Inc.New YorkUSA

Personalised recommendations