SPARO: Selective Attention for Robust and Compositional Transformer Encodings for Vision | SpringerLink
Skip to main content
Springer Nature Link
Log in

SPARO: Selective Attention for Robust and Compositional Transformer Encodings for Vision

  • Conference paper
  • First Online:
Computer Vision – ECCV 2024 (ECCV 2024)

Part of the book series: Lecture Notes in Computer Science ((LNCS,volume 15124))

Included in the following conference series:

  • 127 Accesses

Abstract

Selective attention helps us focus on task-relevant aspects in the constant flood of our sensory input. This constraint in our perception allows us to robustly generalize under distractions and to new compositions of perceivable concepts. Transformers employ a similar notion of attention in their architecture, but representation learning models with transformer backbones like CLIP and DINO often fail to demonstrate robustness and compositionality. We highlight a missing architectural prior: unlike human perception, transformer encodings do not separately attend over individual concepts. In response, we propose Sparo, a read-out mechanism that partitions encodings into separately-attended slots, each produced by a single attention head. Using Sparo with CLIP imparts an inductive bias that the vision and text modalities are different views of a shared compositional world with the same corresponding concepts. Using Sparo, we demonstrate improvements on downstream recognition, robustness, retrieval, and compositionality benchmarks with CLIP (up to \(+14\%\) for ImageNet, \(+4\%\) for SugarCrepe), and on nearest neighbors and linear probe for ImageNet with DINO (\(+3\%\) each). We also showcase a powerful ability to intervene and select individual Sparo concepts to further improve downstream task performance (up from \(+4\%\) to \(+9\%\) for SugarCrepe) and use this ability to study the robustness of Sparo ’s representation structure. Finally, we provide insights through ablation experiments and visualization of learned concepts.

This is a preview of subscription content, log in via an institution to check access.

Access this chapter

Log in via an institution

Subscribe and save

Springer+ Basic
¥17,985 /Month
  • Get 10 units per month
  • Download Article/Chapter or eBook
  • 1 Unit = 1 Article or 1 Chapter
  • Cancel anytime
Subscribe now

Buy Now

Chapter
JPY 3498
Price includes VAT (Japan)
  • Available as PDF
  • Read on any device
  • Instant download
  • Own it forever
eBook
JPY 8007
Price includes VAT (Japan)
  • Available as EPUB and PDF
  • Read on any device
  • Instant download
  • Own it forever
Softcover Book
JPY 10009
Price includes VAT (Japan)
  • Compact, lightweight edition
  • Dispatched in 3 to 5 business days
  • Free shipping worldwide - see info

Tax calculation will be finalised at checkout

Purchases are for personal use only

Institutional subscriptions

Similar content being viewed by others

Notes

  1. 1.

    Source code: https://github.com/ankitkv/sparo-clip.

References

  1. Aydemir, G., Xie, W., Güney, F.: Self-supervised object-centric learning for videos. arXiv preprint arXiv:2310.06907 (2023)

  2. Ba, J.L., Kiros, J.R., Hinton, G.E.: Layer normalization. arXiv preprint arXiv:1607.06450 (2016)

  3. Bahdanau, D., Cho, K., Bengio, Y.: Neural machine translation by jointly learning to align and translate. arXiv preprint arXiv:1409.0473 (2014)

  4. Barbu, A., et al.: ObjectNet: a large-scale bias-controlled dataset for pushing the limits of object recognition models. In: Advances in Neural Information Processing Systems, vol. 32 (2019)

    Google Scholar 

  5. Beattie, C., et al.: DeepMind lab. arXiv preprint arXiv:1612.03801 (2016)

  6. Boff, K.R., Kaufman, L., Thomas, J.P.: Handbook of Perception and Human Performance, vol. 1. Wiley, New York (1986)

    Google Scholar 

  7. Brady, J., Zimmermann, R.S., Sharma, Y., Schölkopf, B., von Kügelgen, J., Brendel, W.: Provably learning object-centric representations. arXiv preprint arXiv:2305.14229 (2023)

  8. Burgess, C.P., et al.: MoNet: unsupervised scene decomposition and representation. arXiv preprint arXiv:1901.11390 (2019)

  9. Caron, M., et al.: Emerging properties in self-supervised vision transformers. In: Proceedings of the IEEE/CVF International Conference on Computer Vision, pp. 9650–9660 (2021)

    Google Scholar 

  10. Chang, H.S., Sun, R.Y., Ricci, K., McCallum, A.: Multi-CLS BERT: an efficient alternative to traditional ensembling. arXiv preprint arXiv:2210.05043 (2022)

  11. Changpinyo, S., Sharma, P., Ding, N., Soricut, R.: Conceptual 12M: pushing web-scale image-text pre-training to recognize long-tail visual concepts. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp. 3558–3568 (2021)

    Google Scholar 

  12. Chen, X., Xie, S., He, K.: An empirical study of training self-supervised vision transformers. arXiv preprint arXiv:2104.02057 (2021)

  13. Chen, Y., et al.: Revisiting multimodal representation in contrastive learning: from patch and token embeddings to finite discrete tokens. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp. 15095–15104 (2023)

    Google Scholar 

  14. Cheng, G., Han, J., Lu, X.: Remote sensing image scene classification: benchmark and state of the art. Proc. IEEE 105(10), 1865–1883 (2017)

    Article  Google Scholar 

  15. Cimpoi, M., Maji, S., Kokkinos, I., Mohamed, S., Vedaldi, A.: Describing textures in the wild. In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pp. 3606–3613 (2014)

    Google Scholar 

  16. Colby, C.L., Goldberg, M.E.: Space and attention in parietal cortex. Annu. Rev. Neurosci. 22(1), 319–349 (1999)

    Article  Google Scholar 

  17. Deng, J., Dong, W., Socher, R., Li, L.J., Li, K., Fei-Fei, L.: ImageNet: a large-scale hierarchical image database. In: 2009 IEEE Conference on Computer Vision and Pattern Recognition, pp. 248–255. IEEE (2009)

    Google Scholar 

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

    Google Scholar 

  19. Dosovitskiy, A., et al.: An image is worth 16 \(\times \) 16 words: transformers for image recognition at scale. arXiv preprint arXiv:2010.11929 (2020)

  20. Dugas, E., Jared, Jorge, Cukierski, W.: Diabetic retinopathy detection (2015). https://kaggle.com/competitions/diabetic-retinopathy-detection

  21. Engelcke, M., Kosiorek, A.R., Jones, O.P., Posner, I.: Genesis: generative scene inference and sampling with object-centric latent representations. arXiv preprint arXiv:1907.13052 (2019)

  22. Eslami, S., et al.: Attend, infer, repeat: fast scene understanding with generative models. In: Advances in Neural Information Processing Systems, vol. 29 (2016)

    Google Scholar 

  23. Fei-Fei, L., Fergus, R., Perona, P.: One-shot learning of object categories. IEEE Trans. Pattern Anal. Mach. Intell. 28(4), 594–611 (2006)

    Article  Google Scholar 

  24. Geiger, A., Lenz, P., Stiller, C., Urtasun, R.: Vision meets robotics: the KITTI dataset. Int. J. Robot. Res. 32(11), 1231–1237 (2013)

    Article  Google Scholar 

  25. Goyal, A., et al.: Neural production systems. Adv. Neural. Inf. Process. Syst. 34, 25673–25687 (2021)

    Google Scholar 

  26. Goyal, A., et al.: Factorizing declarative and procedural knowledge in structured, dynamical environments. In: International Conference on Learning Representations (2020)

    Google Scholar 

  27. Goyal, A., et al.: Recurrent independent mechanisms. arXiv preprint arXiv:1909.10893 (2019)

  28. Greff, K., et al.: Multi-object representation learning with iterative variational inference. In: International Conference on Machine Learning, pp. 2424–2433. PMLR (2019)

    Google Scholar 

  29. 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, pp. 770–778 (2016)

    Google Scholar 

  30. Helber, P., Bischke, B., Dengel, A., Borth, D.: EuroSat: a novel dataset and deep learning benchmark for land use and land cover classification. IEEE J. Sel. Top. Appl. Earth Observ. Remote Sens. 12(7), 2217–2226 (2019)

    Article  Google Scholar 

  31. Hendrycks, D., et al.: The many faces of robustness: a critical analysis of out-of-distribution generalization. In: Proceedings of the IEEE/CVF International Conference on Computer Vision, pp. 8340–8349 (2021)

    Google Scholar 

  32. Hendrycks, D., Zhao, K., Basart, S., Steinhardt, J., Song, D.: Natural adversarial examples. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp. 15262–15271 (2021)

    Google Scholar 

  33. Hodosh, M., Young, P., Hockenmaier, J.: Framing image description as a ranking task: data, models and evaluation metrics. J. Arti. Intell. Res. 47, 853–899 (2013)

    MathSciNet  Google Scholar 

  34. Hsieh, C.Y., Zhang, J., Ma, Z., Kembhavi, A., Krishna, R.: SugarCrepe: fixing hackable benchmarks for vision-language compositionality. Adv. Neural Inf. Process. Syst. (2023)

    Google Scholar 

  35. Ilharco, G., et al.: OpenCLIP (2021). https://github.com/mlfoundations/open_clip, https://doi.org/10.5281/zenodo.5143773

  36. Johnson, J., Hariharan, B., van der Maaten, L., Fei-Fei, L., Lawrence Zitnick, C., Girshick, R.: CLEVR: a diagnostic dataset for compositional language and elementary visual reasoning. In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pp. 2901–2910 (2017)

    Google Scholar 

  37. Krizhevsky, A., Hinton, G., et al.: Learning multiple layers of features from tiny images. Technical report (2009)

    Google Scholar 

  38. LAION-AI: CLIP_benchmark open-source project (2022). https://github.com/LAION-AI/CLIP_benchmark

  39. Lavoie, S., et al.: Simplicial embeddings in self-supervised learning and downstream classification. In: International Conference on Learning Representations (2023). https://openreview.net/forum?id=RWtGreRpovS

  40. LeCun, Y., Huang, F.J., Bottou, L.: Learning methods for generic object recognition with invariance to pose and lighting. In: Proceedings of the 2004 IEEE Computer Society Conference on Computer Vision and Pattern Recognition, CVPR 2004, vol. 2, pp. II–104. IEEE (2004)

    Google Scholar 

  41. Lee, J., Lee, Y., Kim, J., Kosiorek, A., Choi, S., Teh, Y.W.: Set transformer: a framework for attention-based permutation-invariant neural networks. In: International Conference on Machine Learning, pp. 3744–3753. PMLR (2019)

    Google Scholar 

  42. 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_48

    Chapter  Google Scholar 

  43. Locatello, F., et al.: Object-centric learning with slot attention. Adv. Neural. Inf. Process. Syst. 33, 11525–11538 (2020)

    Google Scholar 

  44. Loshchilov, I., Hutter, F.: Decoupled weight decay regularization. arXiv preprint arXiv:1711.05101 (2017)

  45. Ma, Z., Hong, J., Gul, M.O., Gandhi, M., Gao, I., Krishna, R.: CREPE: can vision-language foundation models reason compositionally? In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp. 10910–10921 (2023)

    Google Scholar 

  46. Mansouri, A., Hartford, J., Zhang, Y., Bengio, Y.: Object centric architectures enable efficient causal representation learning. In: The Twelfth International Conference on Learning Representations (2024). https://openreview.net/forum?id=r9FsiXZxZt

  47. Martinez, A.: Putting spatial attention on the map: timing and localization of stimulus selection processes in striate and extrastriate visual areas. Vision. Res. 41, 1437–1457 (2001)

    Article  Google Scholar 

  48. Martins, A., Astudillo, R.: From softmax to sparsemax: a sparse model of attention and multi-label classification. In: International Conference on Machine Learning, pp. 1614–1623. PMLR (2016)

    Google Scholar 

  49. Matthey, L., Higgins, I., Hassabis, D., Lerchner, A.: dSprites: disentanglement testing sprites dataset (2017). https://github.com/deepmind/dsprites-dataset/

  50. Meta Research: DINO open-source repository (2021). https://github.com/facebookresearch/dino

  51. Nilsback, M.E., Zisserman, A.: Automated flower classification over a large number of classes. In: 2008 Sixth Indian Conference on Computer Vision, Graphics & Image Processing, pp. 722–729. IEEE (2008)

    Google Scholar 

  52. O’Connor, D.H., Fukui, M.M., Pinsk, M.A., Kastner, S.: Attention modulates responses in the human lateral geniculate nucleus. Nat. Neurosci. 5(11), 1203–1209 (2002)

    Article  Google Scholar 

  53. Parkhi, O.M., Vedaldi, A., Zisserman, A., Jawahar, C.: Cats and dogs. In: 2012 IEEE Conference on Computer Vision and Pattern Recognition, pp. 3498–3505. IEEE (2012)

    Google Scholar 

  54. Paszke, A., et al.: PyTorch: an imperative style, high-performance deep learning library. Adv. Neural Inf. Process. Syst. 32 (2019)

    Google Scholar 

  55. Qian, R., Ding, S., Liu, X., Lin, D.: Semantics meets temporal correspondence: self-supervised object-centric learning in videos. In: Proceedings of the IEEE/CVF International Conference on Computer Vision, pp. 16675–16687 (2023)

    Google Scholar 

  56. Radford, A., et al.: Learning transferable visual models from natural language supervision. In: International Conference on Machine Learning, pp. 8748–8763. PMLR (2021)

    Google Scholar 

  57. Radford, A., Wu, J., Child, R., Luan, D., Amodei, D., Sutskever, I., et al.: Language models are unsupervised multitask learners. OpenAI blog 1(8), 9 (2019)

    Google Scholar 

  58. Ray, A., Radenovic, F., Dubey, A., Plummer, B.A., Krishna, R., Saenko, K.: COLA: a benchmark for compositional text-to-image retrieval. In: Thirty-seventh Conference on Neural Information Processing Systems Datasets and Benchmarks Track (2023)

    Google Scholar 

  59. Recht, B., Roelofs, R., Schmidt, L., Shankar, V.: Do ImageNet classifiers generalize to ImageNet? In: International Conference on Machine Learning, pp. 5389–5400. PMLR (2019)

    Google Scholar 

  60. Schuhmann, C., et al.: LAION-400M: open dataset of CLIP-filtered 400 million image-text pairs. arXiv preprint arXiv:2111.02114 (2021)

  61. Scott, W.A.: Cognitive complexity and cognitive flexibility. Sociometry 405–414 (1962)

    Google Scholar 

  62. Seitzer, M., et al.: Bridging the gap to real-world object-centric learning. arXiv preprint arXiv:2209.14860 (2022)

  63. Sharma, P., Ding, N., Goodman, S., Soricut, R.: Conceptual captions: a cleaned, hypernymed, image alt-text dataset for automatic image captioning. In: Proceedings of the 56th Annual Meeting of the Association for Computational Linguistics, ACL 2018, Melbourne, Australia, 15–20 July 2018, Volume 1: Long Papers, pp. 2556–2565. Association for Computational Linguistics (2018)

    Google Scholar 

  64. Shazeer, N.: Fast transformer decoding: one write-head is all you need. arXiv preprint arXiv:1911.02150 (2019)

  65. Thrush, T., et al.: WinoGround: probing vision and language models for visio-linguistic compositionality. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp. 5238–5248 (2022)

    Google Scholar 

  66. Treisman, A.M., Gelade, G.: A feature-integration theory of attention. Cogn. Psychol. 12(1), 97–136 (1980)

    Article  Google Scholar 

  67. UW RAIVN Lab: SugarCrepe open-source repository (2023). https://github.com/RAIVNLab/sugar-crepe

  68. Vaswani, A., et al.: Attention is all you need. Adv. Neural Inf. Process. Syst. 30 (2017)

    Google Scholar 

  69. Veeling, B.S., Linmans, J., Winkens, J., Cohen, T., Welling, M.: Rotation equivariant CNNs for digital pathology. In: Frangi, A.F., Schnabel, J.A., Davatzikos, C., Alberola-López, C., Fichtinger, G. (eds.) MICCAI 2018. LNCS, vol. 11071, pp. 210–218. Springer, Cham (2018). https://doi.org/10.1007/978-3-030-00934-2_24

    Chapter  Google Scholar 

  70. Wang, H., Ge, S., Lipton, Z., Xing, E.P.: Learning robust global representations by penalizing local predictive power. Adv. Neural Inf. Process. Syst. 10506–10518 (2019)

    Google Scholar 

  71. Xiao, J., Hays, J., Ehinger, K.A., Oliva, A., Torralba, A.: Sun database: large-scale scene recognition from abbey to zoo. In: 2010 IEEE Computer Society Conference on Computer Vision and Pattern Recognition, pp. 3485–3492. IEEE (2010)

    Google Scholar 

  72. Xu, L., Ouyang, W., Bennamoun, M., Boussaid, F., Xu, D.: Multi-class token transformer for weakly supervised semantic segmentation. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp. 4310–4319 (2022)

    Google Scholar 

  73. Young, P., Lai, A., Hodosh, M., Hockenmaier, J.: From image descriptions to visual denotations: new similarity metrics for semantic inference over event descriptions. Trans. Assoc. Comput. Linguist. 2, 67–78 (2014)

    Article  Google Scholar 

  74. Yu, J., Wang, Z., Vasudevan, V., Yeung, L., Seyedhosseini, M., Wu, Y.: CoCa: contrastive captioners are image-text foundation models. arXiv preprint arXiv:2205.01917 (2022)

  75. Yuksekgonul, M., Bianchi, F., Kalluri, P., Jurafsky, D., Zou, J.: When and why vision-language models behave like bags-of-words, and what to do about it? In: The Eleventh International Conference on Learning Representations (2023)

    Google Scholar 

  76. Yuval, N.: Reading digits in natural images with unsupervised feature learning. In: Proceedings of the NIPS Workshop on Deep Learning and Unsupervised Feature Learning (2011)

    Google Scholar 

  77. Zadaianchuk, A., Kleindessner, M., Zhu, Y., Locatello, F., Brox, T.: Unsupervised semantic segmentation with self-supervised object-centric representations. arXiv preprint arXiv:2207.05027 (2022)

  78. Zhai, X., et al.: A large-scale study of representation learning with the visual task adaptation benchmark. arXiv preprint arXiv:1910.04867 (2019)

  79. Zhang, Y., Hare, J., Prugel-Bennett, A.: Deep set prediction networks. Adv. Neural Inf. Process. Syst. 32 (2019)

    Google Scholar 

  80. Zhao, T., et al.: VL-checklist: evaluating pre-trained vision-language models with objects, attributes and relations. arXiv preprint arXiv:2207.00221 (2022)

Download references

Acknowledgments

This research was funded by Sony and enabled in part by compute resources provided by the Digital Research Alliance of Canada, Mila, and Sony.

Author information

Authors and Affiliations

  1. Mila, Université de Montréal, Montreal, Canada

    Ankit Vani, Samuel Lavoie & Aaron Courville

  2. Sony AI, Stuttgart, Germany

    Bac Nguyen

  3. University of Washington, Allen Institute for Artificial Intelligence, Seattle, USA

    Ranjay Krishna

  4. CIFAR AI Chair, Montreal, Canada

    Aaron Courville

Authors
  1. Ankit Vani
    View author publications

    You can also search for this author in PubMed Google Scholar

  2. Bac Nguyen
    View author publications

    You can also search for this author in PubMed Google Scholar

  3. Samuel Lavoie
    View author publications

    You can also search for this author in PubMed Google Scholar

  4. Ranjay Krishna
    View author publications

    You can also search for this author in PubMed Google Scholar

  5. Aaron Courville
    View author publications

    You can also search for this author in PubMed Google Scholar

Corresponding author

Correspondence to Ankit Vani .

Editor information

Editors and Affiliations

  1. University of Birmingham, Birmingham, UK

    Aleš Leonardis

  2. University of Trento, Trento, Italy

    Elisa Ricci

  3. Technical University of Darmstadt, Darmstadt, Germany

    Stefan Roth

  4. Princeton University, Princeton, NJ, USA

    Olga Russakovsky

  5. Czech Technical University in Prague, Prague, Czech Republic

    Torsten Sattler

  6. École des Ponts ParisTech, Marne-la-Vallée, France

    Gül Varol

1 Electronic supplementary material

Below is the link to the electronic supplementary material.

Supplementary material 1 (pdf 7191 KB)

Rights and permissions

Reprints and permissions

Copyright information

© 2025 The Author(s), under exclusive license to Springer Nature Switzerland AG

About this paper

Check for updates. Verify currency and authenticity via CrossMark

Cite this paper

Vani, A., Nguyen, B., Lavoie, S., Krishna, R., Courville, A. (2025). SPARO: Selective Attention for Robust and Compositional Transformer Encodings for Vision. In: Leonardis, A., Ricci, E., Roth, S., Russakovsky, O., Sattler, T., Varol, G. (eds) Computer Vision – ECCV 2024. ECCV 2024. Lecture Notes in Computer Science, vol 15124. Springer, Cham. https://doi.org/10.1007/978-3-031-72848-8_14

Download citation

  • DOI: https://doi.org/10.1007/978-3-031-72848-8_14

  • Published:

  • Publisher Name: Springer, Cham

  • Print ISBN: 978-3-031-72847-1

  • Online ISBN: 978-3-031-72848-8

  • eBook Packages: Computer ScienceComputer Science (R0)

Publish with us

Policies and ethics

Search

Navigation