{"type":"ScholarlyArticle","title":"Handling Black Swan Events in Deep Learning with Diversely Extrapolated Neural Networks","authors":[{"name":"Maxime Wabartha","slug":"maxime-wabartha","given":"Maxime","family":"Wabartha","affiliation":"McGill University; Mila"},{"name":"Audrey Durand","slug":"audrey-durand","given":"Audrey","family":"Durand","affiliation":"Université Laval; Mila"},{"name":"Vincent François-Lavet","slug":"vincent-francois-lavet","given":"Vincent","family":"François-Lavet","affiliation":"Université de Louvain"},{"name":"Joelle Pineau","slug":"joelle-pineau","given":"Joelle","family":"Pineau","affiliation":"McGill University; Facebook AI Research; Mila"}],"abstract":"By virtue of their expressive power, neural networks (NNs) are well suited to fitting large, complex datasets, yet they are also known to \r\nproduce similar predictions for points outside the training distribution.\r\nAs such, they are, like humans, under the influence of the Black Swan theory: models tend to be extremely \"surprised\" by rare events, leading to potentially disastrous consequences, while justifying these same events in hindsight.\r\nTo avoid this pitfall, we introduce DENN, an ensemble approach building a set of Diversely Extrapolated Neural Networks that fits the training data and is able to generalize more diversely when extrapolating to novel data points.\r\nThis leads DENN to output highly uncertain predictions for unexpected inputs.\r\nWe achieve this by adding a diversity term in the loss function used to train the model, computed at specific inputs.\r\nWe first illustrate the usefulness of the method on a low-dimensional regression problem. \r\nThen, we show how the loss can be adapted to tackle anomaly detection during classification, as well as safe imitation learning problems.","doi":"10.24963/ijcai.2020/296","url":"https://ijcai.org/proceedings/2020/296","pdf":"https://ijcai.org/proceedings/2020/0296.pdf","markdown":"https://ijcai.org/proceedings/2020/296.md","bibtex":"https://ijcai.org/proceedings/2020/bibtex/296","volume":{"slug":"2020","title":"Proceedings of the Twenty-Ninth International Joint Conference on Artificial Intelligence","short_title":"IJCAI 2020","url":"https://ijcai.org/proceedings/2020/","year":2020,"location":"Yokohama","dates":"January 2021 (postponed from 11-17 July 2020)","editors":"Christian Bessiere","isbn":"978-0-9992411-6-5","publisher":"International Joint Conferences on Artificial Intelligence Organization","doi_prefix":"10.24963/ijcai.2020"},"section":"Machine Learning","track":"Main track","number":296,"pages":{"first":2140,"last":2147},"keywords":["Machine Learning: Deep Learning","Uncertainty in AI: Uncertainty Representations","Machine Learning: Ensemble Methods"],"videos":[{"label":"Video","url":"https://videos.ijcai20.org/25240/file_25240.m3u8","length_sec":289,"provider":"hls"}],"references":[{"position":1,"raw":"[Abbasi et al., 2019] Mahdieh Abbasi, Changjian Shui, Are- zoo Rajabi, Christian Gagne, and Rakesh Bobba. Toward metrics for differentiating out-of-distribution sets. In Ad- vances in Neural Information Processing Systems Work- shop on Safety and Robustness in Decision Making, 2019.","title":"Toward metrics for differentiating out-of-distribution sets","authors":"Abbasi","year":"2019","venue":"Advances in Neural Information Processing Systems Workshop on Safety and Robustness in Decision Making","doi":null,"url":null,"internal_url":null},{"position":2,"raw":"[Borgwardt et al., 2006] Karsten M Borgwardt, Arthur Gret- ton, Malte J Rasch, Hans-Peter Kriegel, Bernhard Schölkopf, and Alex J Smola. Integrating structured bio- logical data by kernel maximum mean discrepancy. Bioin- formatics, 22(14):e49-e57, 2006.","title":"Integrating structured biological data by kernel maximum mean discrepancy","authors":"Karsten M Borgwardt, Arthur Gretton, J Malte, Hans-Peter Rasch, Bernhard Kriegel, Alex J Schölkopf, Smola","year":"2006","venue":"Bioinformatics","doi":null,"url":null,"internal_url":null},{"position":3,"raw":"[Cai et al., 2006] Hongmin Cai, Xiaoyin Xu, Ju Lu, Jeff W Lichtman, SP Yung, and Stephen TC Wong. Repul- sive force based snake model to segment and track neu- ronal axons in 3d microscopy image stacks. NeuroImage, 32(4):1608-1620, 2006.","title":"Repulsive force based snake model to segment and track neuronal axons in 3d microscopy image stacks","authors":"Cai","year":"2006","venue":"NeuroImage","doi":null,"url":null,"internal_url":null},{"position":4,"raw":"[Cohen et al., 2017] Gregory Cohen, Saeed Afshar, Jonathan Tapson, and André van Schaik. EMNIST: an extension of MNIST to handwritten letters. CoRR, abs/1702.05373, 2017.","title":"EMNIST: an extension of MNIST to handwritten letters","authors":"Gregory Cohen, Saeed Afshar, Jonathan Tapson, André Van Schaik","year":"2017","venue":null,"doi":null,"url":null,"internal_url":null},{"position":5,"raw":"[Cybenko, 1989] George Cybenko. Approximation by su- perpositions of a sigmoidal function. Mathematics of con- trol, signals and systems, 2(4):303-314, 1989.","title":"Approximation by superpositions of a sigmoidal function","authors":"George Cybenko","year":"1989","venue":"Mathematics of control, signals and systems","doi":null,"url":null,"internal_url":null},{"position":6,"raw":"[Efron, 1982] Bradley Efron. The jackknife, the bootstrap, and other resampling plans. Siam, 1982.","title":"The jackknife, the bootstrap, and other resampling plans","authors":"Bradley Efron","year":"1982","venue":null,"doi":null,"url":null,"internal_url":null},{"position":7,"raw":"[Frankle and Carbin, 2019] Jonathan Frankle and Michael Carbin. The lottery ticket hypothesis: Finding sparse, trainable neural networks. In International Conference on Learning Representations, ICLR 2019, New Orleans, LA, USA, May 6-9, 2019, Conference Track Proceedings, 2019.","title":"The lottery ticket hypothesis: Finding sparse, trainable neural networks","authors":"Carbin Frankle, Jonathan Frankle, Michael Carbin","year":"2019","venue":"International Conference on Learning Representations, ICLR 2019","doi":null,"url":null,"internal_url":null},{"position":8,"raw":"[Gal and Ghahramani, 2016] Yarin Gal and Zoubin Ghahra- mani. Dropout as a bayesian approximation: Representing model uncertainty in deep learning. In International Con- ference on Machine Learning, pages 1050-1059, 2016.","title":"Yarin Gal and Zoubin Ghahramani. Dropout as a bayesian approximation: Representing model uncertainty in deep learning","authors":"Ghahramani Gal","year":"2016","venue":"International Conference on Machine Learning","doi":null,"url":null,"internal_url":null},{"position":9,"raw":"[Gal, 2016] Yarin Gal. Uncertainty in deep learning. PhD thesis, University of Cambridge, 2016.","title":"Uncertainty in deep learning","authors":"Yarin Gal","year":"2016","venue":null,"doi":null,"url":null,"internal_url":null},{"position":10,"raw":"[Hafner et al., 2019] Danijar Hafner, Dustin Tran, Timo- thy P. Lillicrap, Alex Irpan, and James Davidson. Noise contrastive priors for functional uncertainty. In Proceed- ings of the Thirty-Fifth Conference on Uncertainty in Ar- tificial Intelligence, UAI 2019, Tel Aviv, Israel, July 22-25, 2019, 2019.","title":"Noise contrastive priors for functional uncertainty","authors":"Hafner","year":"2019","venue":"Proceedings of the Thirty-Fifth Conference on Uncertainty in Artificial Intelligence, UAI 2019","doi":null,"url":null,"internal_url":null},{"position":11,"raw":"[Han et al., 2015] Song Han, Jeff Pool, John Tran, and William Dally. Learning both weights and connections for efficient neural network. In Advances in neural informa- tion processing systems, pages 1135-1143, 2015.","title":"Learning both weights and connections for efficient neural network","authors":"Han","year":"2015","venue":"Advances in neural information processing systems","doi":null,"url":null,"internal_url":null},{"position":12,"raw":"[Hendrycks et al., 2019] Dan Hendrycks, Mantas Mazeika, and Thomas G Dietterich. Deep anomaly detection with outlier exposure. In International Conference on Learning Representations, ICLR 2019, New Orleans, LA, USA, May 6-9, 2019, Conference Track Proceedings, 2019.","title":"Deep anomaly detection with outlier exposure","authors":"Dan Hendrycks, Mantas Mazeika, Thomas G Dietterich","year":"2019","venue":"International Conference on Learning Representations, ICLR 2019","doi":null,"url":null,"internal_url":null},{"position":13,"raw":"[Hong et al., 2018] Zhang-Wei Hong, Tzu-Yun Shann, Shih- Yang Su, Yi-Hsiang Chang, Tsu-Jui Fu, and Chun-Yi Lee. Diversity-driven exploration strategy for deep reinforce- ment learning. In Advances in Neural Information Pro- cessing Systems, pages 10510-10521, 2018.","title":"Diversity-driven exploration strategy for deep reinforcement learning","authors":"Zhang-Wei Hong, Tzu-Yun Shann, Shih-Yang Su, Yi-Hsiang Chang, Tsu-Jui Fu, Chun-Yi Lee","year":"2018","venue":"Advances in Neural Information Processing Systems","doi":null,"url":null,"internal_url":null},{"position":14,"raw":"[Jaynes, 1957] Edwin T Jaynes. Information theory and sta- tistical mechanics. Physical review, 1957.","title":"Information theory and statistical mechanics","authors":"T Edwin, Jaynes","year":"1957","venue":"Physical review","doi":null,"url":null,"internal_url":null},{"position":15,"raw":"[Kim and Pineau, 2013] Beomjoon Kim and Joelle Pineau. Maximum mean discrepancy imitation learning. In Robotics: Science and systems, 2013.","title":"Maximum mean discrepancy imitation learning","authors":"Pineau Kim, Beomjoon Kim, Joelle Pineau","year":"2013","venue":"Robotics: Science and systems","doi":null,"url":null,"internal_url":null},{"position":16,"raw":"[Kingma and Welling, 2014] Diederik P. Kingma and Max Welling. Auto-encoding variational bayes. In Inter- national Conference on Learning Representations, ICLR 2014, Banff, AB, Canada, April 14-16, 2014, Conference Track Proceedings, 2014.","title":"Auto-encoding variational bayes","authors":"P Diederik, Max Kingma, Welling","year":"2014","venue":"International Conference on Learning Representations, ICLR 2014","doi":null,"url":null,"internal_url":null},{"position":17,"raw":"[Lakshminarayanan et al., 2017] Balaji Lakshminarayanan, Alexander Pritzel, and Charles Blundell. Simple and scal- able predictive uncertainty estimation using deep ensem- bles. In Advances in Neural Information Processing Sys- tems, pages 6402-6413, 2017.","title":"Simple and scalable predictive uncertainty estimation using deep ensembles","authors":"Alexander Balaji Lakshminarayanan, Charles Pritzel, Blundell","year":"2017","venue":"Advances in Neural Information Processing Systems","doi":null,"url":null,"internal_url":null},{"position":18,"raw":"[Lee and Chung, 2020] Jisoo Lee and Sae-Young Chung. Robust training with ensemble consensus. In International Conference on Learning Representations, ICLR 2020, Ad- dis Ababa, Ethiopia , April 26-May 1, 2020, Conference Track Proceedings, 2020.","title":"Robust training with ensemble consensus","authors":"Lee, Chung ;, Jisoo Lee, Sae-Young Chung","year":"2020","venue":"International Conference on Learning Representations, ICLR 2020, Addis","doi":null,"url":null,"internal_url":null},{"position":19,"raw":"[Lee et al., 2018] Kimin Lee, Honglak Lee, Kibok Lee, and Jinwoo Shin. Training confidence-calibrated classifiers for detecting out-of-distribution samples. In International Conference on Learning Representations, ICLR 2018, Vancouver, BC, Canada, April 30-May 3, 2018, Confer- ence Track Proceedings, 2018.","title":"Training confidence-calibrated classifiers for detecting out-of-distribution samples","authors":"Lee","year":"2018","venue":"International Conference on Learning Representations, ICLR 2018","doi":null,"url":null,"internal_url":null},{"position":20,"raw":"[Li et al., 2018] Chunyuan Li, Heerad Farkhoor, Rosanne Liu, and Jason Yosinski. Measuring the intrinsic dimen- sion of objective landscapes. In International Conference on Learning Representations, ICLR 2018, Vancouver, BC, Canada, April 30-May 3, 2018, Conference Track Pro- ceedings, 2018.","title":"Measuring the intrinsic dimension of objective landscapes","authors":"Chunyuan Li, Heerad Farkhoor, Rosanne Liu, Jason Yosinski","year":"2018","venue":"International Conference on Learning Representations, ICLR 2018","doi":null,"url":null,"internal_url":null},{"position":21,"raw":"[Malinin and Gales, 2018] Andrey Malinin and Mark Gales. Predictive uncertainty estimation via prior networks. In Advances in Neural Information Processing Systems, pages 7047-7058, 2018.","title":"Predictive uncertainty estimation via prior networks","authors":"Andrey Malinin, Mark Gales","year":"2018","venue":"Advances in Neural Information Processing Systems","doi":null,"url":null,"internal_url":null},{"position":22,"raw":"[Osband et al., 2016] Ian Osband, Charles Blundell, Alexan- der Pritzel, and Benjamin Van Roy. Deep exploration via bootstrapped dqn. In Advances in Neural Information Pro- cessing Systems, pages 4026-4034, 2016.","title":"Deep exploration via bootstrapped dqn","authors":"Ian Osband, Charles Blundell, Alexander Pritzel, Benjamin Van, Roy","year":"2016","venue":"Advances in Neural Information Processing Systems","doi":null,"url":null,"internal_url":null},{"position":23,"raw":"[Osband et al., 2018] Ian Osband, John Aslanides, and Albin Cassirer. Randomized prior functions for deep reinforce- ment learning. In Advances in Neural Information Pro- cessing Systems, pages 8626-8638, 2018.","title":"Randomized prior functions for deep reinforcement learning","authors":"Ian Osband, John Aslanides, Albin Cassirer","year":"2018","venue":"Advances in Neural Information Processing Systems","doi":null,"url":null,"internal_url":null},{"position":24,"raw":"[Pearce et al., 2018] Tim Pearce, Nicolas Anastassacos, Mo- hamed Zaki, and Andy Neely. Bayesian inference with anchored ensembles of neural networks, and application to reinforcement learning. In International Conference on Machine Learning Workshop on Exploration in Reinforce- ment Learning, 2018.","title":"Bayesian inference with anchored ensembles of neural networks, and application to reinforcement learning","authors":"Tim Pearce, Nicolas Anastassacos, Mohamed Zaki, Andy Neely","year":"2018","venue":"International Conference on Machine Learning Workshop on Exploration in Reinforcement Learning","doi":null,"url":null,"internal_url":null},{"position":25,"raw":"[Ross et al., 2011] Stéphane Ross, Geoffrey Gordon, and Drew Bagnell. A reduction of imitation learning and struc- tured prediction to no-regret online learning. In Interna- tional Conference on Artificial Intelligence and Statistics, pages 627-635, 2011.","title":"A reduction of imitation learning and structured prediction to no-regret online learning","authors":"Stéphane Ross, Geoffrey Gordon, Drew Bagnell","year":"2011","venue":"International Conference on Artificial Intelligence and Statistics","doi":null,"url":null,"internal_url":null},{"position":26,"raw":"[Schulman et al., 2017] John Schulman, Filip Wolski, Pra- fulla Dhariwal, Alec Radford, and Oleg Klimov. Proximal policy optimization algorithms. arXiv:1707.06347, 2017.","title":"Proximal policy optimization algorithms","authors":"John Schulman, Filip Wolski, Prafulla Dhariwal, Alec Radford, Oleg Klimov","year":"2017","venue":null,"doi":null,"url":null,"internal_url":null},{"position":27,"raw":"[Sensoy et al., 2020] Murat Sensoy, Lance Kaplan, Federico Cerutti, and Maryam Saleki. Uncertainty-aware deep clas- sifiers using generative models. In Thirty-Fourth AAAI Conference on Artificial Intelligence, 2020.","title":"Uncertainty-aware deep classifiers using generative models","authors":"Murat Sensoy, Lance Kaplan, Federico Cerutti, Maryam Saleki","year":"2020","venue":"Thirty-Fourth AAAI Conference on Artificial Intelligence","doi":null,"url":null,"internal_url":null},{"position":28,"raw":"[Taleb, 2007] Nassim Nicholas Taleb. The black swan: The impact of the highly improbable, volume 2. Random house, 2007.","title":"The black swan: The impact of the highly improbable","authors":"Nassim Nicholas, Taleb","year":"2007","venue":"Random house","doi":null,"url":null,"internal_url":null},{"position":29,"raw":"[Todorov et al., 2012] Emanuel Todorov, Tom Erez, and Yu- val Tassa. Mujoco: A physics engine for model-based control. In 2012 IEEE/RSJ International Conference on Intelligent Robots and Systems, pages 5026-5033. IEEE, 2012.","title":"Mujoco: A physics engine for model-based control","authors":"Emanuel Todorov, Tom Erez, Yuval Tassa","year":"2012","venue":"IEEE/RSJ International Conference on Intelligent Robots and Systems","doi":null,"url":null,"internal_url":null},{"position":30,"raw":"[Tran et al., 2020] Linh Tran, Bastiaan S. Veeling, Kevin Roth, Jakub Swiatkowski, Joshua V. Dillon, Jasper Snoek, Stephan Mandt, Tim Salimans, Sebastian Nowozin, and Rodolphe Jenatton. Hydra: Preserving ensemble diversity for model distillation. arXiv:2001.04694, 2020.","title":"Hydra: Preserving ensemble diversity for model distillation","authors":"Linh Tran, Bastiaan S Veeling, Kevin Roth, Jakub Swiatkowski, Joshua V Dillon, Jasper Snoek, Stephan Mandt, Tim Salimans, Sebastian Nowozin, Rodolphe Jenatton","year":"2020","venue":null,"doi":null,"url":null,"internal_url":null},{"position":31,"raw":"[Vadakkepat et al., 2000] Prahlad Vadakkepat, Kay Chen Tan, and Wang Ming-Liang. Evolutionary artificial poten- tial fields and their application in real time robot path plan- ning. In Congress on evolutionary computation. CEC00 (Cat. No. 00TH8512), volume 1, pages 256-263. IEEE, 2000.","title":"Evolutionary artificial potential fields and their application in real time robot path plan-ning","authors":"Prahlad Vadakkepat, Kay Chen Tan, Wang Ming-Liang","year":"2000","venue":"Congress on evolutionary computation","doi":null,"url":null,"internal_url":null}],"license_note":"Copyright International Joint Conferences on Artificial Intelligence."}