Advertisement

Learning What Makes a Difference from Counterfactual Examples and Gradient Supervision

Conference paper
  • 727 Downloads
Part of the Lecture Notes in Computer Science book series (LNCS, volume 12355)

Abstract

One of the primary challenges limiting the applicability of deep learning is its susceptibility to learning spurious correlations rather than the underlying mechanisms of the task of interest. The resulting failure to generalise cannot be addressed by simply using more data from the same distribution. We propose an auxiliary training objective that improves the generalization capabilities of neural networks by leveraging an overlooked supervisory signal found in existing datasets. We use pairs of minimally-different examples with different labels, a.k.a counterfactual or contrasting examples, which provide a signal indicative of the underlying causal structure of the task. We show that such pairs can be identified in a number of existing datasets in computer vision (visual question answering, multi-label image classification) and natural language processing (sentiment analysis, natural language inference). The new training objective orients the gradient of a model’s decision function with pairs of counterfactual examples. Models trained with this technique demonstrate improved performance on out-of-distribution test sets.

Notes

Acknowledgements

This material is based on research sponsored by Air Force Research Laboratory and DARPA under agreement number FA8750-19-2-0501. The U.S. Government is authorized to reproduce and distribute reprints for Governmental purposes notwithstanding any copyright notation thereon.

Supplementary material

504449_1_En_34_MOESM1_ESM.pdf (9.1 mb)
Supplementary material 1 (pdf 9291 KB)

References

  1. 1.
    Abbasnejad, E., Shi, Q., van den Hengel, A., Liu, L.: A generative adversarial density estimator. In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (2019)Google Scholar
  2. 2.
    Abbasnejad, E., Teney, D., Parvaneh, A., Shi, J., van den Hengel, A.: Counterfactual vision and language learning. In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (2020)Google Scholar
  3. 3.
    Abbasnejad, E., Wu, Q., Shi, Q., van den Hengel, A.: What’s to know? Uncertainty as a guide to asking goal-oriented questions. In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (2019)Google Scholar
  4. 4.
    Agarwal, V., Shetty, R., Fritz, M.: Towards causal VQA: revealing and reducing spurious correlations by invariant and covariant semantic editing. arXiv preprint arXiv:1912.07538 (2019)
  5. 5.
    Agrawal, A., Batra, D., Parikh, D., Kembhavi, A.: Don’t just assume; look and answer: overcoming priors for visual question answering. In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pp. 4971–4980 (2018)Google Scholar
  6. 6.
    Anderson, P., et al.: Bottom-up and top-down attention for image captioning and VQA. In: Proceedings of the CVPR (2018)Google Scholar
  7. 7.
    Anderson, P., et al.: Vision-and-language navigation: interpreting visually-grounded navigation instructions in real environments. In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pp. 3674–3683 (2018)Google Scholar
  8. 8.
    Antol, S., et al.: VQA: visual question answering. In: Proceedings of the IEEE Conference on Computer Vision (2015)Google Scholar
  9. 9.
    Arjovsky, M., Bottou, L., Gulrajani, I., Lopez-Paz, D.: Invariant risk minimization. arXiv preprint arXiv:1907.02893 (2019)
  10. 10.
    Baradel, F., Neverova, N., Mille, J., Mori, G., Wolf, C.: CoPhy: Counterfactual learning of physical dynamics. arXiv preprint arXiv:1909.12000 (2019)
  11. 11.
    Barbu, A., et al.: ObjectNet: a large-scale bias-controlled dataset for pushing the limits of object recognition models. In: Proceedings of the Advances in Neural Information Processing Systems, pp. 9448–9458 (2019)Google Scholar
  12. 12.
    Bartolo, M., Roberts, A., Welbl, J., Riedel, S., Stenetorp, P.: Beat the AI: investigating adversarial human annotations for reading comprehension. arXiv preprint arXiv:2002.00293 (2020)
  13. 13.
    Bowman, S.R., Angeli, G., Potts, C., Manning, C.D.: A large annotated corpus for learning natural language inference. In: Proceedings of the Conference on Empirical Methods in Natural Language Processing (2015)Google Scholar
  14. 14.
    Cadene, R., Dancette, C., Ben-younes, H., Cord, M., Parikh, D.: RUBi: reducing unimodal biases in visual question answering. arXiv preprint arXiv:1906.10169 (2019)
  15. 15.
    Camburu, O.M., Rocktäschel, T., Lukasiewicz, T., Blunsom, P.: e-SNLI: natural language inference with natural language explanations. In: Proceedings of the Advances in Neural Information Processing Systems, pp. 9539–9549 (2018)Google Scholar
  16. 16.
    Chen, M., D’Arcy, M., Liu, A., Fernandez, J., Downey, D.: CODAH: an adversarially-authored question answering dataset for common sense. In: Proceedings of the 3rd Workshop on Evaluating Vector Space Representations for NLP, pp. 63–69 (2019)Google Scholar
  17. 17.
    Chen, X., et al.: Microsoft COCO captions: data collection and evaluation server. arXiv preprint arXiv:1504.00325 (2015)
  18. 18.
    Clark, C., Yatskar, M., Zettlemoyer, L.: Don’t take the easy way out: ensemble based methods for avoiding known dataset biases. arXiv preprint arXiv:1909.03683 (2019)
  19. 19.
    Teney, D., Abbasnejad, E., van den Hengel, A.: On incorporating semantic prior knowledge in deep learning through embedding-space constraints. arXiv preprint arXiv:1909.13471 (2019)
  20. 20.
    Teney, D., Abbasnejad, E., van den Hengel, A.: Unshuffling data for improved generalization. arXiv preprint arXiv:2002.11894 (2020)
  21. 21.
    Das, A., Agrawal, H., Zitnick, C.L., Parikh, D., Batra, D.: Human attention in visual question answering: do humans and deep networks look at the same regions? In: Proceedings of the Conference on Empirical Methods in Natural Language Processing (2016)Google Scholar
  22. 22.
    Das, A., et al.: Visual dialog. In: Proceedings of the CVPR (2017)Google Scholar
  23. 23.
    Devlin, J., Chang, M.W., Lee, K., Toutanova, K.: BERT: pre-training of deep bidirectional transformers for language understanding. arXiv preprint arXiv:1810.04805 (2018)
  24. 24.
    Feng, S., Wallace, E., Boyd-Graber, J.: Misleading failures of partial-input baselines. arXiv preprint arXiv:1905.05778 (2019)
  25. 25.
    Ganin, Y., et al.: Domain-adversarial training of neural networks. J. Mach. Learn. Res. 17, 1–35 (2016)MathSciNetzbMATHGoogle Scholar
  26. 26.
    Goodfellow, I.J., Shlens, J., Szegedy, C.: Explaining and harnessing adversarial examples. arXiv preprint arXiv:1412.6572 (2014)
  27. 27.
    Goyal, Y., Khot, T., Summers-Stay, D., Batra, D., Parikh, D.: Making the V in VQA matter: elevating the role of image understanding in visual question answering. arXiv preprint arXiv:1612.00837 (2016)
  28. 28.
    Goyal, Y., Khot, T., Summers-Stay, D., Batra, D., Parikh, D.: Making the V in VQA matter: elevating the role of image understanding in visual question answering. In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pp. 6904–6913 (2017)Google Scholar
  29. 29.
    Grand, G., Belinkov, Y.: Adversarial regularization for visual question answering: strengths, shortcomings, and side effects. arXiv preprint arXiv:1906.08430 (2019)
  30. 30.
    Guo, Y., Cheng, Z., Nie, L., Liu, Y., Wang, Y., Kankanhalli, M.: Quantifying and alleviating the language prior problem in visual question answering. arXiv preprint arXiv:1905.04877 (2019)
  31. 31.
    He, K., Zhang, X., Ren, S., Sun, J.: Deep residual learning for image recognition. In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (2016)Google Scholar
  32. 32.
    Heinze-Deml, C., Meinshausen, N.: Conditional variance penalties and domain shift robustness. arXiv preprint arXiv:1710.11469 (2017)
  33. 33.
    Hendricks, L.A., Burns, K., Saenko, K., Darrell, T., Rohrbach, A.: Women also snowboard: overcoming bias in captioning models. In: Ferrari, V., Hebert, M., Sminchisescu, C., Weiss, Y. (eds.) ECCV 2018. LNCS, vol. 11207, pp. 793–811. Springer, Cham (2018).  https://doi.org/10.1007/978-3-030-01219-9_47CrossRefGoogle Scholar
  34. 34.
    Iyyer, M., Wieting, J., Gimpel, K., Zettlemoyer, L.: Adversarial example generation with syntactically controlled paraphrase networks. arXiv preprint arXiv:1804.06059 (2018)
  35. 35.
    Jakubovitz, D., Giryes, R.: Improving DNN robustness to adversarial attacks using Jacobian regularization. In: Ferrari, V., Hebert, M., Sminchisescu, C., Weiss, Y. (eds.) ECCV 2018. LNCS, vol. 11216, pp. 525–541. Springer, Cham (2018).  https://doi.org/10.1007/978-3-030-01258-8_32CrossRefGoogle Scholar
  36. 36.
    Jia, R., Liang, P.: Adversarial examples for evaluating reading comprehension systems. arXiv preprint arXiv:1707.07328 (2017)
  37. 37.
    Kaushik, D., Hovy, E., Lipton, Z.C.: Learning the difference that makes a difference with counterfactually-augmented data. arXiv preprint arXiv:1909.12434 (2019)
  38. 38.
    Li, Y., Cohn, T., Baldwin, T.: Learning robust representations of text. In: Proceedings of the Conference on Empirical Methods in Natural Language Processing, pp. 1979–1985. Association for Computational Linguistics (2016)Google Scholar
  39. 39.
    Lin, T.-Y., et al.: Microsoft COCO: common objects in context. In: Fleet, D., Pajdla, T., Schiele, B., Tuytelaars, T. (eds.) ECCV 2014. LNCS, vol. 8693, pp. 740–755. Springer, Cham (2014).  https://doi.org/10.1007/978-3-319-10602-1_48CrossRefGoogle Scholar
  40. 40.
    Liu, C., Mao, J., Sha, F., Yuille, A.: Attention correctness in neural image captioning. In: Proceedings of the Conference on AAAI (2017)Google Scholar
  41. 41.
    Liu, Y., et al.: CBNet: a novel composite backbone network architecture for object detection. arXiv preprint arXiv:1909.03625 (2019)
  42. 42.
    Maas, A.L., Daly, R.E., Pham, P.T., Huang, D., Ng, A.Y., Potts, C.: Learning word vectors for sentiment analysis. In: Proceedings of the 49th Annual Meeting of the Association for Computational Linguistics: Human Language Technologies, vol. 1, pp. 142–150. Association for Computational Linguistics (2011)Google Scholar
  43. 43.
    Mahabadi, R.K., Henderson, J.: Simple but effective techniques to reduce biases. arXiv preprint arXiv:1909.06321 (2019)
  44. 44.
    Mitchell, T.M.: The need for biases in learning generalizations. Department of Computer Science, Laboratory for Computer Science Research (1980)Google Scholar
  45. 45.
    Miyato, T., Dai, A.M., Goodfellow, I.: Adversarial training methods for semi-supervised text classification. arXiv preprint arXiv:1605.07725 (2016)
  46. 46.
    Moosavi-Dezfooli, S.M., Fawzi, A., Fawzi, O., Frossard, P.: Universal adversarial perturbations. In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pp. 1765–1773 (2017)Google Scholar
  47. 47.
    Nguyen, A., Yosinski, J., Clune, J.: Deep neural networks are easily fooled: high confidence predictions for unrecognizable images. In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pp. 427–436 (2015)Google Scholar
  48. 48.
    Ni, J., Li, J., McAuley, J.: Justifying recommendations using distantly-labeled reviews and fine-grained aspects. In: Proceedings of the Conference on Empirical Methods in Natural Language Processing, pp. 188–197. Association for Computational Linguistics (2019)Google Scholar
  49. 49.
    Nie, Y., Williams, A., Dinan, E., Bansal, M., Weston, J., Kiela, D.: Adversarial NLI: a new benchmark for natural language understanding. arXiv preprint arXiv:1910.14599 (2019)
  50. 50.
    Park, D.H., Darrell, T., Rohrbach, A.: Robust change captioning. In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pp. 4624–4633 (2019)Google Scholar
  51. 51.
    Park, D.H., Darrell, T., Rohrbach, A.: Viewpoint invariant change captioning. arXiv preprint arXiv:1901.02527 (2019)
  52. 52.
    Didelez, V., Pigeot, I.: Judea Pearl: Causality: models, reasoning, and inference. Politische Vierteljahresschrift 42(2), 313–315 (2001).  https://doi.org/10.1007/s11615-001-0048-3CrossRefGoogle Scholar
  53. 53.
    Peters, J., Bühlmann, P., Meinshausen, N.: Causal inference by using invariant prediction: identification and confidence intervals. J. Roy. Stat. Soc.: Ser. B (Stat. Methodol.) 78, 947–1012 (2016)MathSciNetCrossRefGoogle Scholar
  54. 54.
    Qiao, T., Dong, J., Xu, D.: Exploring human-like attention supervision in visual question answering. In: Proceedings of the Conference on AAAI (2018)Google Scholar
  55. 55.
    Ramakrishnan, S., Agrawal, A., Lee, S.: Overcoming language priors in visual question answering with adversarial regularization. In: Proceedings of the Advances in Neural Information Processing Systems, pp. 1541–1551 (2018)Google Scholar
  56. 56.
    Ribeiro, M.T., Singh, S., Guestrin, C.: ”Why should i trust you ?” Explaining the predictions of any classifier. In: Proceedings of the ACM SIGKDD International Conference on Knowledge Discovery and Data Mining (2016)Google Scholar
  57. 57.
    Rojas-Carulla, M., Schölkopf, B., Turner, R., Peters, J.: Invariant models for causal transfer learning. J. Mach. Learn. Res. 19, 1309–1342 (2018)MathSciNetzbMATHGoogle Scholar
  58. 58.
    Rosenthal, S., Farra, N., Nakov, P.: SemEval-2017 task 4: sentiment analysis in twitter. In: Proceedings of the 11th International Workshop on Semantic Evaluation (SemEval 2017), pp. 502–518. Association for Computational Linguistics (2017)Google Scholar
  59. 59.
    Selvaraju, R.R., Das, A., Vedantam, R., Cogswell, M., Parikh, D., Batra, D.: Grad-CAM: why did you say that? arXiv preprint arXiv:1611.07450 (2016)
  60. 60.
    Selvaraju, R.R., et al.: Taking a hint: leveraging explanations to make vision and language models more grounded. In: Proceedings of the IEEE International Conference on Computer Vision (2019)Google Scholar
  61. 61.
    Shafahi, A., et al.: Adversarial training for free! In: Proceedings of the Advances in Neural Information Processing Systems, pp. 3353–3364 (2019)Google Scholar
  62. 62.
    Shetty, R.R., Fritz, M., Schiele, B.: Adversarial scene editing: automatic object removal from weak supervision. In: Proceedings of the Advances in Neural Information Processing Systems, pp. 7706–7716 (2018)Google Scholar
  63. 63.
    Su, W., et al.: VL-BERT: pre-training of generic visual-linguistic representations. arXiv preprint arXiv:1908.08530 (2019)
  64. 64.
    Suhr, A., Lewis, M., Yeh, J., Artzi, Y.: A corpus of natural language for visual reasoning. In: Proceedings of the Conference on Association for Computational Linguistics, vol. 2, pp. 217–223 (2017)Google Scholar
  65. 65.
    Suhr, A., Zhou, S., Zhang, A., Zhang, I., Bai, H., Artzi, Y.: A corpus for reasoning about natural language grounded in photographs. arXiv preprint arXiv:1811.00491 (2018)
  66. 66.
    Tan, H., Bansal, M.: LXMERT: learning cross-modality encoder representations from transformers. arXiv preprint arXiv:1908.07490 (2019)
  67. 67.
    Teney, D., Anderson, P., He, X., van den Hengel, A.: Tips and tricks for visual question answering: learnings from the 2017 challenge. In: Proceedings of the CVPR (2018)Google Scholar
  68. 68.
    Teney, D., van den Hengel, A.: Zero-shot visual question answering. arXiv preprint arXiv:1611.05546 (2016)
  69. 69.
    Teney, D., van den Hengel, A.: Actively seeking and learning from live data. In: The IEEE Conference on Computer Vision and Pattern Recognition (CVPR) (2019)Google Scholar
  70. 70.
    Torralba, A., Efros, A.A., et al.: Unbiased look at dataset bias. In: Proceedings of the CVPR, vol. 1, p. 7 (2011)Google Scholar
  71. 71.
    Vapnik, V., Izmailov, R.: Rethinking statistical learning theory: learning using statistical invariants. Mach. Learn. 108, 381–423 (2019)MathSciNetCrossRefGoogle Scholar
  72. 72.
    Vapnik, V.N.: An overview of statistical learning theory. IEEE Trans. Neural Netw. 10(5), 988–999 (1999)CrossRefGoogle Scholar
  73. 73.
    Vo, N., et al.: Composing text and image for image retrieval-an empirical Odyssey. In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pp. 6439–6448 (2019)Google Scholar
  74. 74.
    Wallace, E., Boyd-Graber, J.: Trick me if you can: adversarial writing of trivia challenge questions. In: ACL Student Research Workshop (2018)Google Scholar
  75. 75.
    Wang, T., Zhao, J., Yatskar, M., Chang, K.W., Ordonez, V.: Balanced datasets are not enough: estimating and mitigating gender bias in deep image representations. In: Proceedings of the IEEE International Conference on Computer Vision, pp. 5310–5319 (2019)Google Scholar
  76. 76.
    Williams, A., Nangia, N., Bowman, S.R.: A broad-coverage challenge corpus for sentence understanding through inference. arXiv preprint arXiv:1704.05426 (2017)
  77. 77.
    Woods, W., Chen, J., Teuscher, C.: Adversarial explanations for understanding image classification decisions and improved neural network robustness. Nat. Mach. Intell. 1(11), 508–516 (2019)CrossRefGoogle Scholar
  78. 78.
    Xie, C., Tan, M., Gong, B., Wang, J., Yuille, A., Le, Q.V.: Adversarial examples improve image recognition. arXiv preprint arXiv:1911.09665 (2019)
  79. 79.
    Yang, Z., He, X., Gao, J., Deng, L., Smola, A.: Stacked attention networks for image question answering. In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (2016)Google Scholar
  80. 80.
    Yelp: Yelp dataset challenge. http://www.yelp.com/dataset_challenge
  81. 81.
    Zellers, R., Bisk, Y., Schwartz, R., Choi, Y.: SWAG: a large-scale adversarial dataset for grounded commonsense inference. arXiv preprint arXiv:1808.05326 (2018)
  82. 82.
    Zellers, R., Holtzman, A., Bisk, Y., Farhadi, A., Choi, Y.: HellaSwag: can a machine really finish your sentence? arXiv preprint arXiv:1905.07830 (2019)
  83. 83.
    Zhang, P., Goyal, Y., Summers-Stay, D., Batra, D., Parikh, D.: Yin and Yang: balancing and answering binary visual questions. In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (2016)Google Scholar

Copyright information

© Springer Nature Switzerland AG 2020

Authors and Affiliations

  1. 1.Australian Institute for Machine LearningUniversity of AdelaideAdelaideAustralia

Personalised recommendations