{"type":"ScholarlyArticle","title":"Human-Centric Justification of Machine Learning Predictions","authors":[{"name":"Or Biran","slug":"or-biran","given":"Or","family":"Biran","affiliation":"Columbia University"},{"name":"Kathleen McKeown","slug":"kathleen-mckeown","given":"Kathleen","family":"McKeown","affiliation":"Columbia University"}],"abstract":"Human decision makers in many domains can make use of predictions made by machine learning models in their decision making process, but the usability of these predictions is limited if the human is unable to justify his or her trust in the prediction. We propose a novel approach to producing justifications that is geared towards users without machine learning expertise, focusing on domain knowledge and on human reasoning, and utilizing natural language generation. Through a task-based experiment, we show that our approach significantly helps humans to correctly decide whether or not predictions are accurate, and significantly increases their satisfaction with the justification.","doi":"10.24963/ijcai.2017/202","url":"https://ijcai.org/proceedings/2017/202","pdf":"https://ijcai.org/proceedings/2017/0202.pdf","markdown":"https://ijcai.org/proceedings/2017/202.md","bibtex":"https://ijcai.org/proceedings/2017/bibtex/202","volume":{"slug":"2017","title":"Proceedings of the Twenty-Sixth International Joint Conference on Artificial Intelligence","short_title":"IJCAI 2017","url":"https://ijcai.org/proceedings/2017/","year":2017,"location":"Melbourne, Australia","dates":"19-25 August 2017","editors":"Carles Sierra","isbn":"978-0-9992411-0-3","publisher":"International Joint Conferences on Artificial Intelligence Organization","doi_prefix":"10.24963/ijcai.2017"},"section":"Machine Learning A-R","track":"Main track","number":202,"pages":{"first":1461,"last":1467},"keywords":["Natural Language Processing: Natural Language Generation","Multidisciplinary Topics and Applications: Validation and Verification","Machine Learning: Classification","Machine Learning: Machine Learning"],"videos":[],"references":[{"position":1,"raw":"[Baehrens et al., 2010] David Baehrens, Timon Schroeter, Stefan Harmeling, Motoaki Kawanabe, Katja Hansen, and Klaus-Robert Müller. How to explain individual classifi- cation decisions. JMLR, 11, August 2010.","title":"How to explain individual classification decisions","authors":"David Baehrens, Timon Schroeter, Stefan Harmeling, Motoaki Kawanabe, Katja Hansen, Klaus-Robert Müller","year":"2010","venue":"JMLR","doi":null,"url":null,"internal_url":null},{"position":2,"raw":"[Barzilay and Lapata, 2005] Regina Barzilay and Mirella Lapata. Collective content selection for concept-to-text generation. In HLT/EMNLP, Vancouver, 2005.","title":"Collective content selection for concept-to-text generation","authors":"Lapata Barzilay, Regina Barzilay, Mirella Lapata","year":"2005","venue":"HLT/EMNLP","doi":null,"url":null,"internal_url":null},{"position":3,"raw":"[Barzilay et al., 1998] Regina Barzilay, Daryl Mccullough, Owen Rambow, Jonathan DeCristofaro, Tanya Korelsky, and Benoit Lavoie. A new approach to expert system ex- planations. In International Workshop on NLG, 1998.","title":"A new approach to expert system explanations","authors":"Regina Barzilay, Daryl Mccullough, Owen Rambow, Jonathan Decristofaro, Tanya Korelsky, Benoit Lavoie","year":"1998","venue":"International Workshop on NLG","doi":null,"url":null,"internal_url":null},{"position":4,"raw":"[Bilgic and Mooney, 2005] Mustafa Bilgic and Raymond J. Mooney. Explaining recommendations: Satisfaction vs. promotion. In Workshop on the Next Stage of Recom- mender Systems Research, San Diego, CA, 2005.","title":"Explaining recommendations: Satisfaction vs. promotion","authors":"Mustafa Bilgic, Raymond J Mooney","year":"2005","venue":"Workshop on the Next Stage of Recommender Systems Research","doi":null,"url":null,"internal_url":null},{"position":5,"raw":"[Biran and McKeown, 2015] Or Biran and Kathleen McKe- own. Discourse planning with an n-gram model of rela- tions. In Empirical Methods in Natural Language Pro- cessing (EMNLP), Lisbon, Portugal, 2015.","title":"Or Biran and Kathleen McKeown. Discourse planning with an n-gram model of relations","authors":"Mckeown Biran","year":"2015","venue":"Empirical Methods in Natural Language Processing (EMNLP)","doi":null,"url":null,"internal_url":null},{"position":6,"raw":"[Carrizosa et al., 2006] Emilio Carrizosa, Belen Martin, and Dolores Morales. A column generation approach for sup- port vector machines. Technical report, 2006.","title":"A column generation approach for support vector machines","authors":"Emilio Carrizosa, Belen Martin, Dolores Morales","year":"2006","venue":null,"doi":null,"url":null,"internal_url":null},{"position":7,"raw":"[Druzdzel, 1996] Marek J Druzdzel. Qualitiative verbal ex- planations in bayesian belief networks. AISB QUAR- TERLY, pages 43-54, 1996.","title":"Qualitiative verbal explanations in bayesian belief networks","authors":"J Marek, Druzdzel","year":"1996","venue":"AISB QUAR-TERLY","doi":null,"url":null,"internal_url":null},{"position":8,"raw":"[Feraud and Clerot, 2002] Raphael Feraud and Fabrice Clerot. A methodology to explain neural network classification. Neural Networks, 15(2):237 -246, 2002.","title":"Raphael Feraud and Fabrice Clerot. A methodology to explain neural network classification","authors":"Clerot Feraud","year":"2002","venue":"Neural Networks","doi":null,"url":null,"internal_url":null},{"position":9,"raw":"[Helldin and Riveiro, 2009] Tove Helldin and Maria Riveiro. Explanation methods for bayesian networks: review and application to a maritime scenario. In Skövde Workshop on Information Fusion Topics. University of Skövde, 2009.","title":"Explanation methods for bayesian networks: review and application to a maritime scenario","authors":"Riveiro Helldin, Tove Helldin, Maria Riveiro","year":"2009","venue":"Skövde Workshop on Information Fusion Topics","doi":null,"url":null,"internal_url":null},{"position":10,"raw":"[Herlocker et al., 2000] Jonathan Herlocker, Joseph Kon- stan, and John Riedl. Explaining collaborative filtering recommendations. In Computer Supported Cooperative Work (CSCW), 2000.","title":"Explaining collaborative filtering recommendations","authors":"Jonathan Herlocker, Joseph Konstan, John Riedl","year":"2000","venue":"Computer Supported Cooperative Work (CSCW)","doi":null,"url":null,"internal_url":null},{"position":11,"raw":"[Jakulin et al., 2005] Aleks Jakulin, Martin Možina, Janez Demšar, Ivan Bratko, and Blaž Zupan. Nomograms for visualizing support vector machines. In Knowledge Dis- covery in Data Mining (KDD), 2005.","title":"Nomograms for visualizing support vector machines","authors":"Aleks Jakulin, Martin Možina, Janez Demšar, Ivan Bratko, Blaž Zupan","year":"2005","venue":"Knowledge Discovery in Data Mining (KDD)","doi":null,"url":null,"internal_url":null},{"position":12,"raw":"[Kononenko et al., 2013] Igor Kononenko, Erik Strumbelj, Zoran Bosnic, Darko Pevec, Matjaz Kukar, and Marko Robnik-Sikonja. Explanation and reliability of individual predictions. Informatica (Slovenia), 37(1):41-48, 2013.","title":"Explanation and reliability of individual predictions","authors":"Igor Kononenko, Erik Strumbelj, Zoran Bosnic, Darko Pevec, Matjaz Kukar, Marko Robnik-Sikonja","year":"2013","venue":"Informatica (Slovenia)","doi":null,"url":null,"internal_url":null},{"position":13,"raw":"[Lacave and Díez, 2002] C. Lacave and F. J. Díez. A review of explanation methods for Bayesian networks. Knowl- edge Engineering Review, 17:107-127, 2002.","title":"A review of explanation methods for Bayesian networks","authors":"C Lacave, F J Díez","year":"2002","venue":"Knowledge Engineering Review","doi":null,"url":null,"internal_url":null},{"position":14,"raw":"[Lim and Dey, 2010] Brian Lim and Anind Dey. Toolkit to support intelligibility in context-aware applications. In Ubiquitous Computing (Ubicomp), 2010.","title":"Toolkit to support intelligibility in context-aware applications","authors":"Brian Lim, Anind Dey","year":"2010","venue":"Ubiquitous Computing (Ubicomp)","doi":null,"url":null,"internal_url":null},{"position":15,"raw":"[Lubsen et al., 1978] J. Lubsen, J. Pool, and E. van der Does. A practical device for the application of a diagnostic or prognostic function. Methods of information in medicine, 17(2):127-129, April 1978.","title":"A practical device for the application of a diagnostic or prognostic function","authors":"J Lubsen, J Pool, E Van Der Does","year":"1978","venue":"Methods of information in medicine","doi":null,"url":null,"internal_url":null},{"position":16,"raw":"[Možina et al., 2004] M. Možina, J. Demšar, M. Kattan, and B. Zupan. Nomograms for visualization of naive bayesian classifier. In Principles and Practice of Knowledge Dis- covery in Databases (PKDD), 2004.","title":"Nomograms for visualization of naive bayesian classifier","authors":"M Možina, J Demšar, M Kattan, B Zupan","year":"2004","venue":"Principles and Practice of Knowledge Discovery in Databases (PKDD)","doi":null,"url":null,"internal_url":null},{"position":17,"raw":"[Papadimitriou et al., 2012] A. Papadimitriou, P. Symeoni- dis, and Y. Manolopoulos. A generalized taxonomy of explanations styles for traditional and social recommender systems. Data Min. Knowl. Discov., 24(3):555-583, 2012.","title":"A generalized taxonomy of explanations styles for traditional and social recommender systems","authors":"A Papadimitriou, P Symeonidis, Y Manolopoulos","year":"2012","venue":"Data Min. Knowl. Discov","doi":null,"url":null,"internal_url":null},{"position":18,"raw":"[Prasad et al., 2008] Rashmi Prasad, Nikhil Dinesh, Alan Lee, Eleni Miltsakaki, Livio Robaldo, Aravind Joshi, and Bonnie Webber. The penn discourse treebank 2.0. In Pro- ceedings of LREC, 2008.","title":"The penn discourse treebank 2.0","authors":"Rashmi Prasad, Nikhil Dinesh, Alan Lee, Eleni Miltsakaki, Livio Robaldo, Aravind Joshi, Bonnie Webber","year":"2008","venue":"Proceedings of LREC","doi":null,"url":null,"internal_url":null},{"position":19,"raw":"[Rahwan and Simari, 2009] Iyad Rahwan and Guillermo R. Simari. Argumentation in Artificial Intelligence. Springer Publishing Company, Incorporated, 1st edition, 2009.","title":"Argumentation in Artificial Intelligence","authors":"Simari Rahwan","year":"2009","venue":null,"doi":null,"url":null,"internal_url":null},{"position":20,"raw":"[Reiter, 2007] Ehud Reiter. An architecture for data-to-text systems. In Proceedings of the Eleventh European Work- shop on Natural Language Generation, ENLG '07, 2007.","title":"An architecture for data-to-text systems","authors":"Ehud Reiter, Reiter","year":"2007","venue":"Proceedings of the Eleventh European Workshop on Natural Language Generation, ENLG '07","doi":null,"url":null,"internal_url":null},{"position":21,"raw":"[Ribeiro et al., 2016] Marco Tulio Ribeiro, Sameer Singh, and Carlos Guestrin. \"why should i trust you?\": Explain- ing the predictions of any classifier. In Knowledge Discov- ery and Data Mining (KDD), 2016.","title":"why should i trust you?\": Explaining the predictions of any classifier","authors":"Marco Tulio Ribeiro, Sameer Singh, Carlos Guestrin","year":"2016","venue":"Knowledge Discovery and Data Mining (KDD)","doi":null,"url":null,"internal_url":null},{"position":22,"raw":"[Robnik-Sikonja and Kononenko, 2008] M. Robnik-Sikonja and I. Kononenko. Explaining classifications for individ- ual instances. Knowledge and Data Engineering, IEEE Transactions on, 20(5):589-600, May 2008.","title":"Explaining classifications for individual instances","authors":"M Robnik-Sikonja, I Kononenko","year":"2008","venue":"Knowledge and Data Engineering","doi":null,"url":null,"internal_url":null},{"position":23,"raw":"[Shortliffe and Buchanan, 1975] Edward H Shortliffe and Bruce G Buchanan. A model of inexact reasoning in medicine. Mathematical biosciences, 23(3), 1975.","title":"A model of inexact reasoning in medicine","authors":"H Edward, Bruce G Shortliffe, Buchanan","year":"1975","venue":"Mathematical biosciences","doi":null,"url":null,"internal_url":null},{"position":24,"raw":"[Sinha and Swearingen, 2002] Rashmi Sinha and Kirsten Swearingen. The role of transparency in recommender systems. In CHI EA, 2002.","title":"The role of transparency in recommender systems","authors":"Swearingen Sinha, Rashmi Sinha, Kirsten Swearingen","year":"2002","venue":"CHI EA","doi":null,"url":null,"internal_url":null},{"position":25,"raw":"[Suermondt, 1992] Henri Jacques Suermondt. Explanation in Bayesian Belief Networks. PhD thesis, Stanford, CA, USA, 1992. UMI Order No. GAX92-21673.","title":"Explanation in Bayesian Belief Networks","authors":"Henri Jacques, Suermondt","year":"1992","venue":null,"doi":null,"url":null,"internal_url":null},{"position":26,"raw":"[Swartout, 1983] William R. Swartout. Xplain: A system for creating and explaining expert consulting programs. Artif. Intell., 21(3), September 1983.","title":"Xplain: A system for creating and explaining expert consulting programs","authors":"R William, Swartout","year":"1983","venue":"Artif. Intell","doi":null,"url":null,"internal_url":null},{"position":27,"raw":"[Symeonidis et al., 2009] P. Symeonidis, A. Nanopoulos, and Y. Manolopoulos. Moviexplain: A recommender sys- tem with explanations. In RecSys, 2009.","title":"Moviexplain: A recommender system with explanations","authors":"P Symeonidis, A Nanopoulos, Y Manolopoulos","year":"2009","venue":"RecSys","doi":null,"url":null,"internal_url":null},{"position":28,"raw":"[Szafron et al., 2003] Duane Szafron, Russell Greiner, Paul Lu, David Wishart, Cam Macdonell, John Anvik, Brett Poulin, Zhiyong Lu, and Roman Eisner. Explaining naive bayes classifications. Technical report, 2003.","title":"Explaining naive bayes classifications","authors":"Duane Szafron, Russell Greiner, Paul Lu, David Wishart, Cam Macdonell, John Anvik, Brett Poulin, Zhiyong Lu, Roman Eisner","year":"2003","venue":null,"doi":null,"url":null,"internal_url":null},{"position":29,"raw":"[Timmer et al., 2017] S. Timmer, J. Meyer, H. Prakken, S. Renooij, and B. Verheij. A two-phase method for ex- tracting explanatory arguments from bayesian networks. Int. J. Approx. Reasoning, 80(C):475-494, January 2017.","title":"A two-phase method for extracting explanatory arguments from bayesian networks","authors":"S Timmer, J Meyer, H Prakken, S Renooij, B Verheij","year":"2017","venue":"Int. J. Approx. Reasoning","doi":null,"url":null,"internal_url":null},{"position":30,"raw":"[Tullio et al., 2007] Joe Tullio, Anind Dey, Jason Chalecki, and James Fogarty. How it works: A field study of non- technical users interacting with an intelligent system. In SIGCHI Human Factors in Computing Systems, 2007.","title":"How it works: A field study of nontechnical users interacting with an intelligent system","authors":"Joe Tullio, Anind Dey, Jason Chalecki, James Fogarty","year":"2007","venue":"SIGCHI Human Factors in Computing Systems","doi":null,"url":null,"internal_url":null},{"position":31,"raw":"[Vlek et al., 2016] Charlotte S. Vlek, Henry Prakken, Silja Renooij, and Bart Verheij. A method for explaining bayesian networks for legal evidence with scenarios. Arti- ficial Intelligence and Law, 24(3):285-324, 2016.","title":"A method for explaining bayesian networks for legal evidence with scenarios","authors":"Charlotte S Vlek, Henry Prakken, Silja Renooij, Bart Verheij","year":"2016","venue":"Artificial Intelligence and Law","doi":null,"url":null,"internal_url":null},{"position":32,"raw":"[Yap et al., 2008] Ghim-Eng Yap, Ah-Hwee Tan, and Hwee- Hwa Pang. Explaining inferences in bayesian networks. Applied Intelligence, 29(3):263-278, 2008.","title":"Explaining inferences in bayesian networks","authors":"Yap","year":"2008","venue":"Applied Intelligence","doi":null,"url":null,"internal_url":null}],"license_note":"Copyright International Joint Conferences on Artificial Intelligence."}