Skip to main content

Thank you for visiting You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

Shortcut learning in deep neural networks


Deep learning has triggered the current rise of artificial intelligence and is the workhorse of today’s machine intelligence. Numerous success stories have rapidly spread all over science, industry and society, but its limitations have only recently come into focus. In this Perspective we seek to distil how many of deep learning’s failures can be seen as different symptoms of the same underlying problem: shortcut learning. Shortcuts are decision rules that perform well on standard benchmarks but fail to transfer to more challenging testing conditions, such as real-world scenarios. Related issues are known in comparative psychology, education and linguistics, suggesting that shortcut learning may be a common characteristic of learning systems, biological and artificial alike. Based on these observations, we develop a set of recommendations for model interpretation and benchmarking, highlighting recent advances in machine learning to improve robustness and transferability from the lab to real-world applications.

Access options

Rent or Buy article

Get time limited or full article access on ReadCube.


All prices are NET prices.

Fig. 1: Examples of shortcut learning.
Fig. 2: Toy example of shortcut learning in neural networks.
Fig. 3: Taxonomy of decision rules.
Fig. 4: Humans and DNNs both generalize, but they generalize very differently.

Code availability

Code to reproduce the toy experiment (Fig. 2) is available at:


  1. 1.

    He, K., Zhang, X., Ren, S. & Sun, J. Delving deep into rectifiers: Surpassing human-level performance on ImageNet classification. In Proc. IEEE Int. Conf. Computer Vision 1026–1034 (ACM, 2015).

  2. 2.

    Silver, D. et al. Mastering the game of Go with deep neural networks and tree search. Nature 529, 484–489 (2016).

    Article  Google Scholar 

  3. 3.

    Moravčík, M. et al. Deepstack: expert-level artificial intelligence in heads-up no-limit poker. Science 356, 508–513 (2017).

    MathSciNet  Article  Google Scholar 

  4. 4.

    Rajpurkar, P. et al. CheXNet: radiologist-level pneumonia detection on chest X-rays with deep learning. Preprint at (2017).

  5. 5.

    Devlin, J., Chang, M. W., Lee, K. & Toutanova, K. BERT: pre-training of deep bidirectional transformers for language understanding. In Proc. Annual Conf. North American Chapter of the Association for Computational Linguistics (ACL, 2019).

  6. 6.

    Rolnick, D. et al. Tackling climate change with machine learning. Preprint at (2019).

  7. 7.

    Reichstein, M. et al. Deep learning and process understanding for data-driven earth system science. Nature 566, 195–204 (2019).

    Article  Google Scholar 

  8. 8.

    Szegedy, C. et al. Intriguing properties of neural networks. In Proc. Int. Conf. Learning Representations (ICLR, 2014).

  9. 9.

    Beery, S., Van Horn, G. & Perona, P. Recognition in terra incognita. In European Conf. Computer Vision 456–473 (Springer, 2018).

  10. 10.

    Rosenfeld, A., Zemel, R. & Tsotsos, J. K. The elephant in the room. Preprint at (2018).

  11. 11.

    Heuer, H., Monz, C. & Smeulders, A. W. Generating captions without looking beyond objects. Preprint at (2016).

  12. 12.

    Buolamwini, J. & Gebru, T. Gender shades: Intersectional accuracy disparities in commercial gender classification. In Proc. ACM Fairness Accountability and Transparency 77–91 (PMLR, 2018).

  13. 13.

    Dastin, J. Amazon scraps secret AI recruiting tool that showed bias against women. Reuters (2018).

  14. 14.

    Shane, J. Do neural nets dream of electric sheep? AI Wierdness (2018).

  15. 15.

    Niven, T. & Kao, H.-Y. Probing neural network comprehension of natural language arguments. In Proc. 57th Annual Meeting of the Association of Computational Linguistics 4658–4664 (2019).

  16. 16.

    Jia, R. & Liang, P. Adversarial examples for evaluating reading comprehension systems. Preprint at (2017).

  17. 17.

    Zech, J. R. et al. Variable generalization performance of a deep learning model to detect pneumonia in chest radiographs: a cross-sectional study. PLoS Med. 15, e1002683 (2018).

    Article  Google Scholar 

  18. 18.

    Bickel, S., Bru¨ckner, M. & Scheffer, T. Discriminative learning under covariate shift. J. Mach. Learn. Res. 10, 2137–2155 (2009).

    MathSciNet  MATH  Google Scholar 

  19. 19.

    Schölkopf, B. et al. On causal and anticausal learning. In Proc. Int. Conf. Machine Learning 1255–1262 (ICML, 2012).

  20. 20.

    Torralba, A. & Efros, A. A. Unbiased look at dataset bias. In Proc. IEEE Conf. Computer Vision and Pattern Recognition (IEEE, 2011).

  21. 21.

    Branwen, G. The neural net tank urban legend. (2011).

  22. 22.

    Pfungst, O. Clever Hans (The Horse of Mr. Von Osten): A Contribution to Experimental Animal and Human Psychology (Holt, Rinehart and Winston, 1911).

  23. 23.

    Scouller, K. The influence of assessment method on students’ learning approaches: multiple choice question examination versus assignment essay. Higher Educ. 35, 453–472 (1998).

    Article  Google Scholar 

  24. 24.

    Wichmann, F. A., Drewes, J., Rosas, P. & Gegenfurtner, K. R. Animal detection in natural scenes: critical features revisited. J. Vis. 10, 6 (2010).

    Article  Google Scholar 

  25. 25.

    Ribeiro, M. T., Singh, S. & Guestrin, C. “Why should I trust you?”: Explaining the predictions of any classifier. In Proc. 22nd ACM SIGKDD Int. Conf. Knowledge Discovery and Data Mining 1135–1144 (ACM, 2016).

  26. 26.

    Zhu, Z., Xie, L. & Yuille, A. L. Object recognition with and without objects. In Proc. 26th Int. Joint Conf. Artificial Intelligence 3609–3615 (IJCAI, 2017).

  27. 27.

    Wang, J. et al. Visual concepts and compositional voting. Ann. Math. Sci. Appl. 3, 151–188 (2018).

    MathSciNet  MATH  Google Scholar 

  28. 28.

    Dawson, M., Zisserman, A. & Nellåker, C. From same photo: cheating on visual kinship challenges. In Asian Conf. Computer Vision 654–668 (Springer, 2018).

  29. 29.

    Biederman, I. On the Semantics of a Glance at a Scene (Erlbaum, 1981).

  30. 30.

    Biederman, I., Mezzanotte, R. J. & Rabinowitz, J. C. Scene perception: detecting and judging objects undergoing relational violations. Cogn. Psychol. 14, 143–177 (1982).

    Article  Google Scholar 

  31. 31.

    Oliva, A. & Torralba, A. The role of context in object recognition. Trends Cogn. Sci. 11, 520–527 (2007).

    Article  Google Scholar 

  32. 32.

    Castelhano, M. S. & Heaven, C. Scene context influences without scene gist: eye movements guided by spatial associations in visual search. Psychon. Bull Rev. 18, 890–896 (2011).

    Article  Google Scholar 

  33. 33.

    Jo, J. & Bengio, Y. Measuring the tendency of CNNs to learn surface statistical regularities. Preprint at (2017).

  34. 34.

    Ilyas, A. et al. Adversarial examples are not bugs, they are features. In Proc. Advances NeurIPS 125–136 (NeurIPS, 2019).

  35. 35.

    Wolpert, D. H. & Macready, W. G. No free lunch theorems for optimization. IEEE T. Evolut. Comput. 1, 67–82 (1997).

    Article  Google Scholar 

  36. 36.

    Brendel, W. & Bethge, M. Approximating CNNs with bag-of-local-features models works surprisingly well on ImageNet. In Proc. Int. Conf. Learning Representations (ICLR, 2019).

  37. 37.

    Baker, N., Lu, H., Erlikhman, G. & Kellman, P. J. Deep convolutional networks do not classify based on global object shape. PLoS Comp. Biol. 14, e1006613 (2018).

    Article  Google Scholar 

  38. 38.

    Geirhos, R. et al. ImageNet-trained CNNs are biased towards texture; increasing shape bias improves accuracy and robustness. In Proc. Int. Conf. Learning Representations (ICLR, 2019).

  39. 39.

    Heinze-Deml, C. & Meinshausen, N. Conditional variance penalties and domain shift robustness. Preprint at (2017).

  40. 40.

    Malhotra, G. & Bowers, J. What a difference a pixel makes: an empirical examination of features used by CNNs for categorisation. In Proc. Int. Conf. Learning Representations (ICLR, 2019).

  41. 41.

    Jacobsen, J.-H., Behrmann, J., Zemel, R. & Bethge, M. Excessive invariance causes adversarial vulnerability. In Proc. Int. Conf. Learning Representations (ICLR, 2019).

  42. 42.

    Kamin, L. J. Predictability, surprise, attention, and conditioning. In Symp. Punishment and Averse Behavior (eds Campbell, B. A. & Church, R. M.) 279–296 (Appleton-Century-Crofts, 1969).

  43. 43.

    Dickinson, A. Contemporary Animal Learning Theory Vol. 1 (CUP Archive, 1980).

  44. 44.

    Bouton, M. E. Learning and Behavior: A Contemporary Synthesis (Sinauer Associates, 2007).

  45. 45.

    Nguyen, A., Yosinski, J. & Clune, J. Deep neural networks are easily fooled: high confidence predictions for unrecognizable images. In Proc. IEEE Conf. Computer Vision and Pattern Recognition 427–436 (IEEE, 2015).

  46. 46.

    Hendrycks, D., Zhao, K., Basart, S., Steinhardt, J. & Song, D. Natural adversarial examples. Preprint at (2019).

  47. 47.

    Wang, M. & Deng, W. Deep visual domain adaptation: a survey. Neurocomputing 312, 135–153 (2018).

    Article  Google Scholar 

  48. 48.

    Alcorn, M. A. et al. Strike (with) a pose: neural networks are easily fooled by strange poses of familiar objects. In Proc. IEEE Conf. Computer Vision and Pattern Recognition (IEEE, 2019).

  49. 49.

    Azulay, A. & Weiss, Y. Why do deep convolutional networks generalize so poorly to small image transformations? J. Mach. Learn. Res. 20, 1–25 (2019).

    MathSciNet  MATH  Google Scholar 

  50. 50.

    Dodge, S. & Karam, L. Human and DNN classification performance on images with quality distortions: a comparative study. ACM T. Appl. Perc. 16, 7 (2019).

    Google Scholar 

  51. 51.

    Russakovsky, O. et al. ImageNet large scale visual recognition challenge. Int. J. Comput. Vis. 115, 211–252 (2015).

    MathSciNet  Article  Google Scholar 

  52. 52.

    Gururangan, S. et al. Annotation artifacts in natural language inference data. In Proc. Annual Conf. North American Chapter of the Association for Computational Linguistics (ACL, 2018).

  53. 53.

    Zellers, R., Holtzman, A., Bisk, Y., Farhadi, A. & Choi, Y. HellaSwag: can a machine really finish your sentence? In Proc. 57th Annual Meeting Assocciation of Computational Linguistics 4791–4800 (ACL, 2019).

  54. 54.

    Borowski, J. et al. The notorious difficulty of comparing human and machine perception. In Proc. NeurIPS Shared Representations in Human and Machine Intelligence Workshop (NeurIPS, 2019).

  55. 55.

    Geirhos, R., Meding, K. & Wichmann, F. A. Beyond accuracy: quantifying trial-by-trial behaviour of CNNs and humans by measuring error consistency. Preprint at (2020).

  56. 56.

    Marr, D. Vision: A Computational Investigation into the Human Representation and Processing of Visual Information (W. H. Freeman and Company, 1982).

  57. 57.

    Buckner, C. The Comparative Psychology of Artificial Intelligences (PhilSci Archive, 2019);

  58. 58.

    Morgan, C. L. Introduction to Comparative Psychology (Scribner, 1903).

  59. 59.

    Ghahramani, Z. Panel of workshop on advances in approximate Bayesian inference (AABI) 2017. YouTube (2017).

  60. 60.

    Marton, F. & Säaljö, R. On qualitative differences in learning—II Outcome as a function of the learner’s conception of the task. Br. J. Educ. Psychol. 46, 115–127 (1976).

    Article  Google Scholar 

  61. 61.

    Biggs, J. Individual differences in study processes and the quality of learning outcomes. Higher Educ. 8, 381–394 (1979).

    Article  Google Scholar 

  62. 62.

    Chin, C. & Brown, D. E. Learning in science: a comparison of deep and surface approaches. J. Res. Sci. Teach. 37, 109–138 (2000).

    Article  Google Scholar 

  63. 63.

    Marcus, G. F. Rethinking eliminative connectionism. Cogn. Psychol. 37, 243–282 (1998).

    Article  Google Scholar 

  64. 64.

    Kilbertus, N., Parascandolo, G. & Schölkopf, B. Generalization in anti-causal learning. Preprint at (2018).

  65. 65.

    Marcus, G. Deep learning: a critical appraisal. Preprint at (2018).

  66. 66.

    Lapuschkin, S. et al. Unmasking Clever Hans predictors and assessing what machines really learn. Nat. Commun. 10, 1096 (2019).

    Article  Google Scholar 

  67. 67.

    Lake, B. M., Ullman, T. D., Tenenbaum, J. B. & Gershman, S. J. Building machines that learn and think like people. Behav. Brain Sci. 40, e253 (2017).

    Article  Google Scholar 

  68. 68.

    Chollet, F. The measure of intelligence. Preprint at (2019).

  69. 69.

    Crosby, M., Beyret, B. & Halina, M. The Animal-AI Olympics. Nat. Mach. Int. 1, 257–257 (2019).

    Article  Google Scholar 

  70. 70.

    Juliani, A. et al. Obstacle tower: a generalization challenge in vision, control, and planning. In Proc. 28th Int. Joint Conf. Artificial Intelligence (IJCAI, 2019).

  71. 71.

    Hendrycks, D. & Dietterich, T. Benchmarking neural network robustness to common corruptions and perturbations. In Proc. Int. Conf. Learning Representations (ICLR, 2019).

  72. 72.

    Levesque, H., Davis, E. & Morgenstern, L. The Winograd Schema Challenge. In 13th Int. Conf. Principles of Knowledge Representation and Reasoning (KR, 2012).

  73. 73.

    Trichelair, P., Emami, A., Trischler, A., Suleman, K. & Cheung, J. C. K. How reasonable are common-sense reasoning tasks: a case-study on the Winograd Schema Challenge and SWAG. In Proc. Conf. Empirical Methods in Natural Language Processing and Int. Joint Conf. Natural Language Processing 3373–3378 (ACL, 2019).

  74. 74.

    Zipf, G. K. Human Behavior and the Principle of Least Effort (Addison-Wesley, 1949).

  75. 75.

    Ohala, J. J. The phonetics and phonology of aspects of assimilation. Papers Lab. Phono. 1, 258–275 (1990).

    Article  Google Scholar 

  76. 76.

    Vicentini, A. The economy principle in language. Notes and Observations from early modern English grammars. Mots Palabras Words 3, 37–57 (2003).

    Google Scholar 

  77. 77.

    Sinz, F. H., Pitkow, X., Reimer, J., Bethge, M. & Tolias, A. S. Engineering a less artificial intelligence. Neuron 103, 967–979 (2019).

    Article  Google Scholar 

  78. 78.

    Arpit, D. et al. A closer look at memorization in deep networks. In Proc. Int. Conf. Machine Learning (ICML, 2017).

  79. 79.

    Valle-Perez, G., Camargo, C. Q. & Louis, A. A. Deep learning generalizes because the parameter-function map is biased towards simple functions. In Proc. Int. Conf. Learning Representations (ICLR, 2018).

  80. 80.

    Shah, H., Tamuly, K., Raghunathan, A., Jain, P. & Netrapalli, P. The pitfalls of simplicity bias in neural networks. Preprint at (2020).

  81. 81.

    Kalimeris, D. et al. SGD on neural networks learns functions of increasing complexity. In Proc. Advances NeurIPS 3496–3506 (NeurIPS, 2019).

  82. 82.

    Hermann, K. L. & Lampinen, A. K. What shapes feature representations? exploring datasets, architectures, and training. Preprint at (2020).

  83. 83.

    Richardson, J. Vectors: Aphorisms & Ten-Second Essays (Ausable, 2001).

  84. 84.

    Engstrom, L. et al. A discussion of ‘adversarial examples are not bugs, they are features’. Distill (2019).

  85. 85.

    Barbu, A. et al. ObjectNet: a large-scale bias-controlled dataset for pushing the limits of object recognition models. In Proc. Advances NeurIPS 9448–9458 (NeurIPS, 2019).

  86. 86.

    Li, D., Yang, Y., Song, Y.-Z. & Hospedales, T. M. Deeper, broader and artier domain generalization. In Proc. IEEE Int. Conf. Computer Vision (IEEE, 2017).

  87. 87.

    Qiu, W. & Yuille, A. UnrealCV: connecting computer vision to unreal engine. In European Conf. Computer Vision 909–916 (Springer, 2016).

  88. 88.

    Dosovitskiy, A., Ros, G., Codevilla, F., Lopez, A. & Koltun, V. CARLA: an open urban driving simulator. In Conf. Robot Learning 1–16 (CoRL, 2017).

  89. 89.

    Creager, E. et al. Flexibly fair representation learning by disentanglement. In Proc. Int. Conf. Machine Learning (ICML, 2019).

  90. 90.

    Hays, J. & Efros, A. A. Scene completion using millions of photographs. ACM Trans. Graph. 26, 4 (2007).

    Article  Google Scholar 

  91. 91.

    Hays, J. & Efros, A. A. IM2GPS: estimating geographic information from a single image. In Proc. IEEE Conf. Computer Vision and Pattern Recognition (IEEE, 2008).

  92. 92.

    Poliak, A., Naradowsky, J., Haldar, A., Rudinger, R. & Van Durme, B. Hypothesis only baselines in natural language inference. In Proc. 7th Joint Conf. Lexical and Computational Semantics 180–191 (ACL, 2018).

  93. 93.

    Jasani, B., Girdhar, R. & Ramanan, D. Are we asking the right questions in MovieQA? In Proc. IEEE/CVF Int. Conf. Computer Vision Workshop (IEEE, 2019).

  94. 94.

    Hornik, K., Stinchcombe, M. & White, H. Multilayer feedforward networks are universal approximators. Neural Netw. 2, 359–366 (1989).

    Article  Google Scholar 

  95. 95.

    d’Ascoli, S., Sagun, L., Bruna, J. & Biroli, G. Finding the needle in the haystack with convolutions: on the benefits of architectural bias. In Proc. Advances NeurIPS (NeurIPS, 2019).

  96. 96.

    Ulyanov, D., Vedaldi, A. & Lempitsky, V. Deep image prior. In Proc. IEEE/CVF Conf. Computer Vision and Pattern Recognition 9446–9454 (IEEE, 2018).

  97. 97.

    Vaswani, A. et al. Attention is all you need. In Proc. Advances NeurIPS 5998–6008 (NeurIPS, 2017).

  98. 98.

    Hein, M., Andriushchenko, M. & Bitterwolf, J. Why ReLU networks yield high-confidence predictions far away from the training data and how to mitigate the problem. In Proc. IEEE Conf. Computer Vision and Pattern Recognition 41–50 (IEEE, 2019).

  99. 99.

    Lehman, J. et al. The surprising creativity of digital evolution: a collection of anecdotes from the evolutionary computation and artificial life research communities. Art. Life 26, 274–306 (2020).

    Article  Google Scholar 

  100. 100.

    Madry, A., Makelov, A., Schmidt, L., Tsipras, D. & Vladu, A. Towards deep learning models resistant to adversarial attacks. In Proc. Int. Conf. Learning Representations (ICLR, 2018).

  101. 101.

    Arjovsky, M., Bottou, L., Gulrajani, I. & Lopez-Paz, D. Invariant risk minimization. Preprint at (2019).

  102. 102.

    Wu, L., Zhu, Z. & E, W. Towards understanding generalization of deep learning: perspective of loss landscapes. Preprint at (2017).

  103. 103.

    De Palma, G., Kiani, B. T. & Lloyd, S. Deep neural networks are biased towards simple functions. Preprint at (2018).

  104. 104.

    Valle-Perez, G., Camargo, C. Q. & Louis, A. A. Deep learning generalizes because the parameter-function map is biased towards simple functions. In Proc. Int. Conf. Learning Representations (ICLR, 2019).

  105. 105.

    Sun, K. & Nielsen, F. Lightlike neuromanifolds, Occam’s razor and deep learning. Preprint at (2019).

  106. 106.

    Li, Y., Wei, C. & Ma, T. Towards explaining the regularization effect of initial large learning rate in training neural networks. In Proc. Advances NeurIPS 11674–11685 (NeurIPS, 2019).

  107. 107.

    Bartlett, P. L., Long, P. M., Lugosi, G. & Tsigler, A. Benign overfitting in linear regression. Proc. Natl Acad Sci. USA (2019).

Download references


The authors thank the International Max Planck Research School for Intelligent Systems (IMPRS-IS) for supporting R.G. and C.M.; the Deutsche Forschungsgemeinschaft (DFG, German Research Foundation) for supporting C.M. via grant EC 479/1-1; the Collaborative Research Center (Projektnummer 276693517—SFB 1233: Robust Vision) for supporting M.B. and F.A.W.; the German Federal Ministry of Education and Research through the Tübingen AI Center (FKZ 01IS18039A) for supporting W.B. and M.B.; as well as the Natural Sciences and Engineering Research Council of Canada and the Intelligence Advanced Research Projects Activity (IARPA) via Department of Interior/Interior Business Center (DoI/IBC) contract number D16PC00003 for supporting R.Z. The authors would like to thank J. Borowski, M. Burg, S. Cadena, A. S. Ecker, L. Eisenberg, R. Fleming, I. Fründ, S. Greiner, F. Grießer, S. Keshvari, R. Kessler, D. Klindt, M. Kümmerer, B. Mitzkus, H. Nienborg, J. Rauber, E. Rusak, S. Schneider, L. Schott, T. Sering, Y. Sharma, M. Tangemann, R. Zimmermann and T. Wallis for helpful discussions.

Author information




The project was initiated by R.G. and C.M. and led by R.G. with support from C.M. and J.J.; F.A.W. added the cognitive science and neuroscience connection; M.B. and W.B. reshaped the initial thrust of the perspective and together with R.Z. supervised the machine learning components. The toy experiment was conducted by J.J. with input from R.G. and C.M. Most figures were designed by R.G. and W.B. with input from all other authors. Figure 2 (left) was conceived by M.B. The first draft was written by R.G., J.J. and C.M. with input from F.A.W. All authors contributed to the final version and provided critical revisions from different perspectives.

Corresponding author

Correspondence to Robert Geirhos.

Ethics declarations

Competing interests

The authors declare no competing interests.

Additional information

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary information

Supplementary Information

Supplementary Sections A–C and references

Rights and permissions

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Geirhos, R., Jacobsen, JH., Michaelis, C. et al. Shortcut learning in deep neural networks. Nat Mach Intell 2, 665–673 (2020).

Download citation

Further reading


Quick links

Nature Briefing

Sign up for the Nature Briefing newsletter — what matters in science, free to your inbox daily.

Get the most important science stories of the day, free in your inbox. Sign up for Nature Briefing