Transparent Model Distillation

Model distillation was originally designed to distill knowledge from a large, complex teacher model to a faster, simpler student model without significant loss in prediction accuracy. We investigate model distillation for another goal -- transparency -- investigating if fully-connected neural networks can be distilled into models that are transparent or interpretable in some sense. Our teacher models are multilayer perceptrons, and we try two types of student models: (1) tree-based generalized additive models (GA2Ms), a type of boosted, short tree (2) gradient boosted trees (GBTs). More transparent student models are forthcoming. Our results are not yet conclusive. GA2Ms show some promise for distilling binary classification teachers, but not yet regression. GBTs are not "directly" interpretable but may be promising for regression teachers. GA2M models may provide a computationally viable alternative to additive decomposition methods for global function approximation.

[1]  Jude W. Shavlik,et al.  in Advances in Neural Information Processing , 1996 .

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

[3]  I. Sobola,et al.  Global sensitivity indices for nonlinear mathematical models and their Monte Carlo estimates , 2001 .

[4]  Chong Gu.  Smoothing Spline Anova Models , 2002 .

[5]  Giles Hooker,et al.  Discovering additive structure in black box functions , 2004, KDD.

[6]  Rich Caruana,et al.  Model compression , 2006, KDD '06.

[7]  S. Wood.  Generalized Additive Models: An Introduction with R , 2006 .

[8]  G. Hooker.  Generalized Functional ANOVA Diagnostics for High-Dimensional Functions of Dependent Variables , 2007 .

[9]  Bogdan E. Popescu,et al.  PREDICTIVE LEARNING VIA RULE ENSEMBLES , 2008, 0811.1679.

[10]  S. Wood.  Fast stable restricted maximum likelihood and marginal likelihood estimation of semiparametric generalized linear models , 2011 .

[11]  Johannes Gehrke,et al.  Intelligible models for classification and regression , 2012, KDD.

[12]  Johannes Gehrke,et al.  Accurate intelligible models with pairwise interactions , 2013, KDD.

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

[14]  Nitish Srivastava,et al.  Dropout: a simple way to prevent neural networks from overfitting , 2014, J. Mach. Learn. Res..

[15]  Kevin Leyton-Brown,et al.  An Efficient Approach for Assessing Hyperparameter Importance , 2014, ICML.

[16]  Rich Caruana,et al.  Do Deep Nets Really Need to be Deep? , 2013, NIPS.

[17]  Geoffrey E. Hinton,et al.  Distilling the Knowledge in a Neural Network , 2015, ArXiv.

[18]  Simon N. Wood,et al.  Shape constrained additive models , 2015, Stat. Comput..

[19]  Sergey Ioffe,et al.  Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift , 2015, ICML.

[20]  B. Iooss,et al.  A Review on Global Sensitivity Analysis Methods , 2014, 1404.2405.

[21]  Jimmy Ba,et al.  Adam: A Method for Stochastic Optimization , 2014, ICLR.

[22]  Johannes Gehrke,et al.  Intelligible Models for HealthCare: Predicting Pneumonia Risk and Hospital 30-day Readmission , 2015, KDD.

[23]  Yoshua Bengio,et al.  FitNets: Hints for Thin Deep Nets , 2014, ICLR.

[24]  Oluwasanmi Koyejo,et al.  Examples are not enough, learn to criticize! Criticism for Interpretability , 2016, NIPS.

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

[26]  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).

[27]  Yan Liu,et al.  Interpretable Deep Models for ICU Outcome Prediction , 2016, AMIA.

[28]  Jure Leskovec,et al.  Interpretable & Explorable Approximations of Black Box Models , 2017, ArXiv.

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

[30]  Osbert Bastani,et al.  Interpreting Blackbox Models via Model Extraction , 2017, ArXiv.

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

[32]  Matthew Richardson,et al.  Do Deep Convolutional Nets Really Need to be Deep and Convolutional? , 2016, ICLR.

[33]  Yan Liu,et al.  Detecting Statistical Interactions from Neural Network Weights , 2017, ICLR.

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