Explanatory Interactive Machine Learning

Although interactive learning puts the user into the loop, the learner remains mostly a black box for the user. Understanding the reasons behind predictions and queries is important when assessing how the learner works and, in turn, trust. Consequently, we propose the novel framework of explanatory interactive learning where, in each step, the learner explains its query to the user, and the user interacts by both answering the query and correcting the explanation. We demonstrate that this can boost the predictive and explanatory powers of, and the trust into, the learned model, using text (e.g. SVMs) and image classification (e.g. neural networks) experiments as well as a user study.

[1]  Carlos Guestrin,et al.  Anchors: High-Precision Model-Agnostic Explanations , 2018, AAAI.

[2]  Paolo Viappiani,et al.  Preferences in Interactive Systems: Technical Challenges and Case Studies , 2008, AI Mag..

[3]  Thorsten Joachims,et al.  Coactive Learning , 2015, J. Artif. Intell. Res..

[4]  Corinna Cortes,et al.  Support-Vector Networks , 1995, Machine Learning.

[5]  James Allan,et al.  An interactive algorithm for asking and incorporating feature feedback into support vector machines , 2007, SIGIR.

[6]  Tom Michael Mitchell,et al.  Explanation-based generalization: A unifying view , 1986 .

[7]  Weng-Keen Wong,et al.  Principles of Explanatory Debugging to Personalize Interactive Machine Learning , 2015, IUI.

[8]  Andreas Krause,et al.  Nonmyopic active learning of Gaussian processes: an exploration-exploitation approach , 2007, ICML '07.

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

[10]  Scott Lundberg,et al.  An unexpected unity among methods for interpreting model predictions , 2016, ArXiv.

[11]  A. Thomaz,et al.  Mixed-Initiative Active Learning , 2012 .

[12]  Andrew Slavin Ross,et al.  Right for the Right Reasons: Training Differentiable Models by Constraining their Explanations , 2017, IJCAI.

[13]  Sriraam Natarajan,et al.  Human-Guided Learning for Probabilistic Logic Models , 2018, Front. Robot. AI.

[14]  Thomas G. Dietterich,et al.  Interacting meaningfully with machine learning systems: Three experiments , 2009, Int. J. Hum. Comput. Stud..

[15]  Susan T. Dumais,et al.  Short-Term Satisfaction and Long-Term Coverage: Understanding How Users Tolerate Algorithmic Exploration , 2018, WSDM.

[16]  Luc De Raedt,et al.  Probabilistic Explanation Based Learning , 2007, ECML.

[17]  David Cohn,et al.  Active Learning , 2010, Encyclopedia of Machine Learning.

[18]  Gerald DeJong,et al.  Explanation-Based Learning , 2014, Encyclopedia of Machine Learning and Data Mining.

[19]  Burr Settles,et al.  Closing the Loop: Fast, Interactive Semi-Supervised Annotation With Queries on Features and Instances , 2011, EMNLP.

[20]  Jason Eisner,et al.  Modeling Annotators: A Generative Approach to Learning from Annotator Rationales , 2008, EMNLP.

[21]  Steve Hanneke,et al.  Theory of Disagreement-Based Active Learning , 2014, Found. Trends Mach. Learn..

[22]  Andrew McCallum,et al.  Toward Optimal Active Learning through Monte Carlo Estimation of Error Reduction , 2001, ICML 2001.

[23]  N. Epley,et al.  The mind in the machine: Anthropomorphism increases trust in an autonomous vehicle , 2014 .

[24]  Christine D. Piatko,et al.  Using “Annotator Rationales” to Improve Machine Learning for Text Categorization , 2007, NAACL.

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

[26]  Jeffrey M. Bradshaw,et al.  Trust in Automation , 2013, IEEE Intelligent Systems.

[27]  Zoubin Ghahramani,et al.  Deep Bayesian Active Learning with Image Data , 2017, ICML.

[28]  Carla E. Brodley,et al.  The Constrained Weight Space SVM: Learning with Ranked Features , 2011, ICML.

[29]  Daphne Koller,et al.  Support Vector Machine Active Learning with Applications to Text Classification , 2000, J. Mach. Learn. Res..

[30]  Ning Wang,et al.  Trust calibration within a human-robot team: Comparing automatically generated explanations , 2016, 2016 11th ACM/IEEE International Conference on Human-Robot Interaction (HRI).

[31]  Andrew McCallum,et al.  Active Learning by Labeling Features , 2009, EMNLP.

[32]  Robert Tibshirani,et al.  1-norm Support Vector Machines , 2003, NIPS.

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

[34]  J. Simpson.  Psychological Foundations of Trust , 2007 .

[35]  Hema Raghavan,et al.  Active Learning with Feedback on Features and Instances , 2006, J. Mach. Learn. Res..

[36]  Manali Sharma,et al.  Active Learning with Rationales for Text Classification , 2015, NAACL.

[37]  Andrew McCallum,et al.  Toward Optimal Active Learning through Sampling Estimation of Error Reduction , 2001, ICML.

[38]  R. Nowak,et al.  Upper and Lower Error Bounds for Active Learning , 2006 .

[39]  Foster J. Provost,et al.  A Unified Approach to Active Dual Supervision for Labeling Features and Examples , 2010, ECML/PKDD.

[40]  Li Chen,et al.  Critiquing-based recommenders: survey and emerging trends , 2012, User Modeling and User-Adapted Interaction.

[41]  Thomas G. Dietterich,et al.  Active Imitation Learning via Reduction to I.I.D. Active Learning , 2012, AAAI Fall Symposium: Robots Learning Interactively from Human Teachers.

[42]  Tom M. Mitchell,et al.  Explanation-Based Generalization: A Unifying View , 1986, Machine Learning.

[43]  Maria-Florina Balcan,et al.  The true sample complexity of active learning , 2010, Machine Learning.

[44]  Gideon S. Mann,et al.  Learning from labeled features using generalized expectation criteria , 2008, SIGIR '08.