A Causal Lens for Peeking into Black Box Predictive Models: Predictive Model Interpretation via Causal Attribution

With the increasing adoption of predictive models trained using machine learning across a wide range of high-stakes applications, e.g., health care, security, criminal justice, finance, and education, there is a growing need for effective techniques for explaining such models and their predictions. We aim to address this problem in settings where the predictive model is a black box; That is, we can only observe the response of the model to various inputs, but have no knowledge about the internal structure of the predictive model, its parameters, the objective function, and the algorithm used to optimize the model. We reduce the problem of interpreting a black box predictive model to that of estimating the causal effects of each of the model inputs on the model output, from observations of the model inputs and the corresponding outputs. We estimate the causal effects of model inputs on model output using variants of the Rubin Neyman potential outcomes framework for estimating causal effects from observational data. We show how the resulting causal attribution of responsibility for model output to the different model inputs can be used to interpret the predictive model and to explain its predictions. We present results of experiments that demonstrate the effectiveness of our approach to the interpretation of black box predictive models via causal attribution in the case of deep neural network models trained on one synthetic data set (where the input variables that impact the output variable are known by design) and two real-world data sets: Handwritten digit classification, and Parkinson's disease severity prediction. Because our approach does not require knowledge about the predictive model algorithm and is free of assumptions regarding the black box predictive model except that its input-output responses be observable, it can be applied, in principle, to any black box predictive model.

[1]  Franco Turini,et al.  A Survey of Methods for Explaining Black Box Models , 2018, ACM Comput. Surv..

[2]  Ankur Taly,et al.  Explainable machine learning in deployment , 2019, FAT*.

[3]  Timothy W. Finin,et al.  The need for user models in generating expert system explanation , 1988 .

[4]  Motoaki Kawanabe,et al.  How to Explain Individual Classification Decisions , 2009, J. Mach. Learn. Res..

[5]  J. Friedman Greedy function approximation: A gradient boosting machine. , 2001 .

[6]  S. Leurgans,et al.  Parkinson disease with old-age onset: a comparative study with subjects with middle-age onset. , 2003, Archives of neurology.

[7]  Jürgen Schmidhuber,et al.  Long Short-Term Memory , 1997, Neural Computation.

[8]  Donglin Zeng,et al.  Personalized Dose Finding Using Outcome Weighted Learning , 2016, Journal of the American Statistical Association.

[9]  Fei Wang,et al.  Deep Learning in Medicine-Promise, Progress, and Challenges. , 2019, JAMA internal medicine.

[10]  Fei Wang,et al.  Deep learning for healthcare: review, opportunities and challenges , 2018, Briefings Bioinform..

[11]  Joseph Y. Halpern,et al.  Causes and explanations: A structural-model approach , 2000 .

[12]  Geoffrey E. Hinton,et al.  Distilling a Neural Network Into a Soft Decision Tree , 2017, CEx@AI*IA.

[13]  Emil Pitkin,et al.  Peeking Inside the Black Box: Visualizing Statistical Learning With Plots of Individual Conditional Expectation , 2013, 1309.6392.

[14]  Vineeth N. Balasubramanian,et al.  Neural Network Attributions: A Causal Perspective , 2019, ICML.

[15]  Julapa Jagtiani,et al.  The Roles of Alternative Data and Machine Learning in Fintech Lending: Evidence from the Lendingclub Consumer Platform , 2018, Financial Management.

[16]  Xintao Wu,et al.  Fairness through Equality of Effort , 2019, WWW.

[17]  Sebastian Thrun,et al.  Extracting Rules from Artifical Neural Networks with Distributed Representations , 1994, NIPS.

[18]  Debashis Ghosh,et al.  A Boosting Algorithm for Estimating Generalized Propensity Scores with Continuous Treatments , 2015, Journal of causal inference.

[19]  Max A. Little,et al.  Accurate Telemonitoring of Parkinson's Disease Progression by Noninvasive Speech Tests , 2009, IEEE Transactions on Biomedical Engineering.

[20]  Dumitru Erhan,et al.  The (Un)reliability of saliency methods , 2017, Explainable AI.

[21]  Huan Liu,et al.  Understanding Neural Networks via Rule Extraction , 1995, IJCAI.

[22]  Hod Lipson,et al.  Understanding Neural Networks Through Deep Visualization , 2015, ArXiv.

[23]  Risto Miikkulainen,et al.  GRADE: Machine Learning Support for Graduate Admissions , 2013, AI Mag..

[24]  P. Cochat,et al.  Et al , 2008, Archives de pediatrie : organe officiel de la Societe francaise de pediatrie.

[25]  Rob Fergus,et al.  Visualizing and Understanding Convolutional Networks , 2013, ECCV.

[26]  Brendan J. Frey,et al.  Machine Learning in Genomic Medicine: A Review of Computational Problems and Data Sets , 2016, Proceedings of the IEEE.

[27]  Valerie Tarasuk,et al.  Liberal trade policy and food insecurity across the income distribution: an observational analysis in 132 countries, 2014–17 , 2020, The Lancet Global Health.

[28]  Scott Lundberg,et al.  A Unified Approach to Interpreting Model Predictions , 2017, NIPS.

[29]  Chris Russell,et al.  Explaining Explanations in AI , 2018, FAT.

[30]  Amina Adadi,et al.  Peeking Inside the Black-Box: A Survey on Explainable Artificial Intelligence (XAI) , 2018, IEEE Access.

[31]  Zachary Chase Lipton The mythos of model interpretability , 2016, ACM Queue.

[32]  Vasant Honavar,et al.  Fairness in Algorithmic Decision Making: An Excursion Through the Lens of Causality , 2019, WWW.

[33]  Justin A. Sirignano,et al.  Deep Learning for Mortgage Risk , 2016, Journal of Financial Econometrics.

[34]  Alexander Binder,et al.  On Pixel-Wise Explanations for Non-Linear Classifier Decisions by Layer-Wise Relevance Propagation , 2015, PloS one.

[35]  Jude W. Shavlik,et al.  Extracting refined rules from knowledge-based neural networks , 2004, Machine Learning.

[36]  Uri Shalit,et al.  Learning Representations for Counterfactual Inference , 2016, ICML.

[37]  William J. Clancey,et al.  Explanation in Human-AI Systems: A Literature Meta-Review, Synopsis of Key Ideas and Publications, and Bibliography for Explainable AI , 2019, ArXiv.

[38]  Abhishek Das,et al.  Grad-CAM: Visual Explanations from Deep Networks via Gradient-Based Localization , 2016, 2017 IEEE International Conference on Computer Vision (ICCV).

[39]  Ilya Shpitser,et al.  Deriving Bounds And Inequality Constraints Using Logical Relations Among Counterfactuals , 2020, UAI.

[40]  Trevor Hastie,et al.  Causal Interpretations of Black-Box Models , 2019, Journal of business & economic statistics : a publication of the American Statistical Association.

[41]  Elias Bareinboim,et al.  Equality of Opportunity in Classification: A Causal Approach , 2018, NeurIPS.

[42]  Avanti Shrikumar,et al.  Learning Important Features Through Propagating Activation Differences , 2017, ICML.

[43]  J. Robins,et al.  Marginal Structural Models and Causal Inference in Epidemiology , 2000, Epidemiology.

[44]  Guigang Zhang,et al.  Deep Learning , 2016, Int. J. Semantic Comput..

[45]  Michael Rabadi,et al.  Kernel Methods for Machine Learning , 2015 .

[46]  Matt J. Kusner,et al.  Counterfactual Fairness , 2017, NIPS.

[47]  Anand Singh Jalal,et al.  Suspicious human activity recognition: a review , 2017, Artificial Intelligence Review.

[48]  Joshua D. Angrist,et al.  Identification of Causal Effects Using Instrumental Variables , 1993 .

[49]  Jenna Wiens,et al.  Machine Learning for Healthcare: On the Verge of a Major Shift in Healthcare Epidemiology , 2018, Clinical infectious diseases : an official publication of the Infectious Diseases Society of America.

[50]  Hans-J. Briegel,et al.  Machine learning \& artificial intelligence in the quantum domain , 2017, ArXiv.

[51]  Le Song,et al.  Learning to Explain: An Information-Theoretic Perspective on Model Interpretation , 2018, ICML.

[52]  Daniel W. Davies,et al.  Machine learning for molecular and materials science , 2018, Nature.

[53]  Jenna Burrell,et al.  How the machine ‘thinks’: Understanding opacity in machine learning algorithms , 2016 .

[54]  K. Borgwardt,et al.  Machine Learning in Medicine , 2015, Mach. Learn. under Resour. Constraints Vol. 3.

[55]  Tim Miller,et al.  Explanation in Artificial Intelligence: Insights from the Social Sciences , 2017, Artif. Intell..

[56]  Ilya Shpitser,et al.  Fair Inference on Outcomes , 2017, AAAI.

[57]  J. Zubizarreta Stable Weights that Balance Covariates for Estimation With Incomplete Outcome Data , 2015 .

[58]  Raimo Tuomela,et al.  A Pragmatic Theory of Explanation , 1984 .

[59]  Erik Strumbelj,et al.  Explaining prediction models and individual predictions with feature contributions , 2014, Knowledge and Information Systems.

[60]  M. J. van der Laan,et al.  Statistical Applications in Genetics and Molecular Biology Super Learner , 2010 .

[61]  Andrew Zisserman,et al.  Deep Inside Convolutional Networks: Visualising Image Classification Models and Saliency Maps , 2013, ICLR.

[62]  M. J. van der Laan,et al.  Practice of Epidemiology Improving Propensity Score Estimators ’ Robustness to Model Misspecification Using Super Learner , 2015 .

[63]  Guillermo Sapiro,et al.  A Shared Vision for Machine Learning in Neuroscience , 2018, The Journal of Neuroscience.

[64]  Anders Larrabee Sønderlund,et al.  The efficacy of learning analytics interventions in higher education: A systematic review , 2018, Br. J. Educ. Technol..

[65]  Been Kim,et al.  Towards A Rigorous Science of Interpretable Machine Learning , 2017, 1702.08608.

[66]  Noemi Kreif,et al.  Machine learning in policy evaluation: new tools for causal inference , 2019, Oxford Research Encyclopedia of Economics and Finance.

[67]  Christophe Croux,et al.  Important factors determining Fintech loan default: Evidence from a lendingclub consumer platform , 2020 .

[68]  Andre Esteva,et al.  A guide to deep learning in healthcare , 2019, Nature Medicine.

[69]  G. Imbens,et al.  The Propensity Score with Continuous Treatments , 2005 .

[70]  Atul J Butte,et al.  A call for deep-learning healthcare , 2019, Nature Medicine.

[71]  Charu C. Aggarwal,et al.  On the Surprising Behavior of Distance Metrics in High Dimensional Spaces , 2001, ICDT.

[72]  M. J. Laan,et al.  Targeted Learning: Causal Inference for Observational and Experimental Data , 2011 .

[73]  Pascal Vincent,et al.  Representation Learning: A Review and New Perspectives , 2012, IEEE Transactions on Pattern Analysis and Machine Intelligence.

[74]  Diogo M. Camacho,et al.  Next-Generation Machine Learning for Biological Networks , 2018, Cell.

[75]  Joseph Y. Halpern,et al.  Causes and Explanations: A Structural-Model Approach. Part II: Explanations , 2001, The British Journal for the Philosophy of Science.

[76]  Stefania Albanesi,et al.  Predicting Consumer Default: A Deep Learning Approach , 2019, SSRN Electronic Journal.

[77]  Max Welling,et al.  Causal Effect Inference with Deep Latent-Variable Models , 2017, NIPS 2017.

[78]  Cynthia Rudin,et al.  Interpretable classifiers using rules and Bayesian analysis: Building a better stroke prediction model , 2015, ArXiv.

[79]  Ribana Roscher,et al.  Explainable Machine Learning for Scientific Insights and Discoveries , 2019, IEEE Access.

[80]  Olivier Bachem,et al.  Recent Advances in Autoencoder-Based Representation Learning , 2018, ArXiv.

[81]  Ankur Taly,et al.  Axiomatic Attribution for Deep Networks , 2017, ICML.

[82]  Wesley C. Salmon,et al.  Van Fraassen on Explanation , 1987 .

[83]  Illtyd Trethowan Causality , 1938 .

[84]  Anna Shcherbina,et al.  Not Just a Black Box: Learning Important Features Through Propagating Activation Differences , 2016, ArXiv.

[85]  D. Rubin,et al.  The central role of the propensity score in observational studies for causal effects , 1983 .

[86]  Wesley C. Salmon,et al.  Causality and Explanation , 1998 .

[87]  Charles F. Manski,et al.  Identification for Prediction and Decision , 2008 .

[88]  Suresh Venkatasubramanian,et al.  Problems with Shapley-value-based explanations as feature importance measures , 2020, ICML.

[89]  Vasant Honavar,et al.  Algorithmic Bias in Recidivism Prediction: A Causal Perspective , 2019, AAAI.

[90]  Kosuke Imai,et al.  Causal Inference With General Treatment Regimes , 2004 .

[91]  E Mjolsness,et al.  Machine learning for science: state of the art and future prospects. , 2001, Science.

[92]  Joachim Diederich,et al.  Survey and critique of techniques for extracting rules from trained artificial neural networks , 1995, Knowl. Based Syst..

[93]  Robert Pelzer,et al.  Policing of Terrorism Using Data from Social Media , 2018, European Journal for Security Research.

[94]  J. Hendler,et al.  Amplify scientific discovery with artificial intelligence , 2014, Science.

[95]  L. Shapley A Value for n-person Games , 1988 .

[96]  Carlos Guestrin,et al.  "Why Should I Trust You?": Explaining the Predictions of Any Classifier , 2016, ArXiv.

[97]  Sebastián Ventura,et al.  Educational Data Mining: A Review of the State of the Art , 2010, IEEE Transactions on Systems, Man, and Cybernetics, Part C (Applications and Reviews).

[98]  Markus H. Gross,et al.  Explaining Deep Neural Networks with a Polynomial Time Algorithm for Shapley Values Approximation , 2019, ICML.

[99]  Walter Karlen,et al.  Perfect Match: A Simple Method for Learning Representations For Counterfactual Inference With Neural Networks , 2018, ArXiv.

[100]  J. Woodward,et al.  Scientific Explanation and the Causal Structure of the World , 1988 .

[101]  S. Lipovetsky,et al.  Analysis of regression in game theory approach , 2001 .

[102]  Yair Zick,et al.  Algorithmic Transparency via Quantitative Input Influence: Theory and Experiments with Learning Systems , 2016, 2016 IEEE Symposium on Security and Privacy (SP).

[103]  Wojciech Samek,et al.  Methods for interpreting and understanding deep neural networks , 2017, Digit. Signal Process..

[104]  Chad Hazlett,et al.  Covariate balancing propensity score for a continuous treatment: Application to the efficacy of political advertisements , 2018 .