Papers
Topics
Authors
Recent
Assistant
AI Research Assistant
Well-researched responses based on relevant abstracts and paper content.
Custom Instructions Pro
Preferences or requirements that you'd like Emergent Mind to consider when generating responses.
Gemini 2.5 Flash
Gemini 2.5 Flash 156 tok/s
Gemini 2.5 Pro 44 tok/s Pro
GPT-5 Medium 23 tok/s Pro
GPT-5 High 22 tok/s Pro
GPT-4o 109 tok/s Pro
Kimi K2 168 tok/s Pro
GPT OSS 120B 455 tok/s Pro
Claude Sonnet 4.5 32 tok/s Pro
2000 character limit reached

Equivariance with Learned Canonicalization Functions (2211.06489v3)

Published 11 Nov 2022 in cs.LG and cs.AI

Abstract: Symmetry-based neural networks often constrain the architecture in order to achieve invariance or equivariance to a group of transformations. In this paper, we propose an alternative that avoids this architectural constraint by learning to produce canonical representations of the data. These canonicalization functions can readily be plugged into non-equivariant backbone architectures. We offer explicit ways to implement them for some groups of interest. We show that this approach enjoys universality while providing interpretable insights. Our main hypothesis, supported by our empirical results, is that learning a small neural network to perform canonicalization is better than using predefined heuristics. Our experiments show that learning the canonicalization function is competitive with existing techniques for learning equivariant functions across many tasks, including image classification, $N$-body dynamics prediction, point cloud classification and part segmentation, while being faster across the board.

Definition Search Book Streamline Icon: https://streamlinehq.com
References (72)
  1. Input convex neural networks. In Precup, D. and Teh, Y. W. (eds.), Proceedings of the 34th International Conference on Machine Learning, volume 70 of Proceedings of Machine Learning Research, pp.  146–155. PMLR, 06–11 Aug 2017.
  2. Point convolutional neural networks by extension operators. arXiv preprint arXiv:1803.10091, 2018.
  3. Estimating or propagating gradients through stochastic neurons for conditional computation. arXiv preprint arXiv:1308.3432, 2013.
  4. Learning invariances in neural networks from training data. Advances in neural information processing systems, 33:17605–17616, 2020.
  5. Probabilistic symmetries and invariant neural networks. The Journal of Machine Learning Research, 21(1):3535–3595, 2020.
  6. Smooth and sparse optimal transport. In International conference on artificial intelligence and statistics, pp.  880–889. PMLR, 2018.
  7. Fast differentiable sorting and ranking. In III, H. D. and Singh, A. (eds.), Proceedings of the 37th International Conference on Machine Learning, volume 119 of Proceedings of Machine Learning Research, pp.  950–959. PMLR, 13–18 Jul 2020. URL https://proceedings.mlr.press/v119/blondel20a.html.
  8. Efficient and modular implicit differentiation, 2022. URL https://openreview.net/forum?id=TQ75Md-FqQp.
  9. Symmetry group equivariant architectures for physics. arXiv preprint arXiv:2203.06153, 2022.
  10. Geometric deep learning: Grids, groups, graphs, geodesics, and gauges. arXiv preprint arXiv:2104.13478, 2021.
  11. Mental rotation and the frame of reference in blind and sighted individuals. Perception & Psychophysics, 23(2):117–124, 1978. doi: 10.3758/BF03208291. URL https://doi.org/10.3758/BF03208291.
  12. Equivariant neural networks for inverse problems. Inverse Problems, 37(8):085006, 2021.
  13. ShapeNet: An Information-Rich 3D Model Repository. Technical Report arXiv:1512.03012 [cs.GR], 2015.
  14. Clusternet: Deep hierarchical cluster network with rigorously rotation-invariant representation for point cloud analysis. pp.  4994–5002, 2019.
  15. Group equivariant convolutional networks. In International conference on machine learning, pp. 2990–2999, 2016a.
  16. Group equivariant convolutional networks. pp.  2990–2999, 2016b.
  17. Cox, G. Almost sure uniqueness of a global minimum without convexity. The Annals of Statistics, 48(1):584 – 606, 2020. doi: 10.1214/19-AOS1829. URL https://doi.org/10.1214/19-AOS1829.
  18. Cuturi, M. Sinkhorn distances: Lightspeed computation of optimal transport. In NeurIPS, 2013.
  19. Vector neurons: A general framework for so (3)-equivariant networks. In Proceedings of the IEEE/CVF International Conference on Computer Vision, pp.  12200–12209, 2021.
  20. SE(3) equivariant graph neural networks with complete local frames. In Chaudhuri, K., Jegelka, S., Song, L., Szepesvari, C., Niu, G., and Sabato, S. (eds.), Proceedings of the 39th International Conference on Machine Learning, pp.  5583–5608, 2022. URL https://proceedings.mlr.press/v162/du22e.html.
  21. Learning SO(3) equivariant representations with spherical cnns. pp.  52–68, 2018a.
  22. Polar transformer networks. In International Conference on Learning Representations, 2018b. URL https://openreview.net/forum?id=HktRlUlAZ.
  23. A practical method for constructing equivariant multilayer perceptrons for arbitrary matrix groups. arXiv preprint arXiv:2104.09459, 2021.
  24. Se(3)-transformers: 3d roto-translation equivariant attention networks. In Advances in Neural Information Processing Systems 34 (NeurIPS), 2020.
  25. On differentiating parameterized argmin and argmax problems with application to bi-level optimization. arXiv preprint arXiv:1607.05447, 2016.
  26. Frames of reference and mental imagery. Attention and performance IX, pp.  261–277, 1981.
  27. Batch normalization: Accelerating deep network training by reducing internal covariate shift. In International conference on machine learning, pp. 448–456. PMLR, 2015.
  28. Spatial transformer networks. Advances in neural information processing systems, 28, 2015.
  29. Adam: A method for stochastic optimization. arXiv preprint arXiv:1412.6980, 2014.
  30. Neural relational inference for interacting systems. In International Conference on Machine Learning, pp. 2688–2697. PMLR, 2018.
  31. Roto-translated local coordinate frames for interacting dynamical systems. In Beygelzimer, A., Dauphin, Y., Liang, P., and Vaughan, J. W. (eds.), Advances in Neural Information Processing Systems, 2021. URL https://openreview.net/forum?id=c3RKZas9am.
  32. An empirical evaluation of deep architectures on problems with many factors of variation. In Proceedings of the 24th International Conference on Machine Learning, ICML ’07, pp.  473–480, 2007.
  33. A rotation-invariant framework for deep point cloud analysis. arXiv preprint arXiv:2003.07238, 2020.
  34. PointCNN: Convolution on x-transformed points. pp.  820–830, 2018.
  35. Deep learning 3d shapes using alt-az anisotropic 2-sphere convolution. 2018.
  36. Investigating bi-level optimization for learning and vision from a unified perspective: A survey and beyond. IEEE Transactions on Pattern Analysis and Machine Intelligence, 44(12):10045–10067, 2021.
  37. Lowe, D. G. Distinctive image features from scale-invariant keypoints. International journal of computer vision, 60(2):91–110, 2004.
  38. Equivariant point cloud analysis via learning orientations for message passing. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pp.  18932–18941, June 2022.
  39. Integral invariants for shape matching. IEEE Transactions on pattern analysis and machine intelligence, 28(10):1602–1618, 2006.
  40. Learning latent permutations with gumbel-sinkhorn networks. In International Conference on Learning Representations, 2018. URL https://openreview.net/forum?id=Byt3oJ-0W.
  41. Group equivariant deep reinforcement learning. arXiv preprint arXiv:2007.03437, 2020.
  42. Effective rotation-invariant point cnn with spherical harmonics kernels. In IEEE International Conference on 3D Vision, pp.  47–56, 2019.
  43. Frame averaging for invariant and equivariant network design. In International Conference on Learning Representations, 2022. URL https://openreview.net/forum?id=zIUyj55nXR.
  44. Pointnet: Deep learning on point sets for 3d classification and segmentation. In Proceedings of the IEEE conference on computer vision and pattern recognition, pp.  652–660, 2017a.
  45. Pointnet++: Deep hierarchical feature learning on point sets in a metric space. pp.  5099–5108, 2017b.
  46. Spherical fractal convolutional neural networks for point cloud recognition. pp.  452–460, 2019.
  47. Equivariance through parameter-sharing. In International Conference on Machine Learning, pp. 2892–2901. PMLR, 2017.
  48. Kornia: an open source differentiable computer vision library for pytorch. In Winter Conference on Applications of Computer Vision, 2020. URL https://arxiv.org/pdf/1910.02190.pdf.
  49. E (n) equivariant graph neural networks. arXiv preprint arXiv:2102.09844, 2021.
  50. On universal equivariant set networks. In International Conference on Learning Representations, 2020. URL https://openreview.net/forum?id=HkxTwkrKDB.
  51. Shawe-Taylor, J. Building symmetries into feedforward networks. In 1989 First IEE International Conference on Artificial Neural Networks, (Conf. Publ. No. 313), pp.  158–162, 1989.
  52. Mental rotation of three-dimensional objects. Science, pp.  701–703, 1971.
  53. Sinkhorn, R. A relationship between arbitrary positive matrices and doubly stochastic matrices. The Annals of Mathematical Statistics, 35(2):876–879, 1964. ISSN 00034851. URL http://www.jstor.org/stable/2238545.
  54. Finding symmetry breaking order parameters with euclidean neural networks. Phys. Rev. Research, 3:L012002, Jan 2021. doi: 10.1103/PhysRevResearch.3.L012002. URL https://link.aps.org/doi/10.1103/PhysRevResearch.3.L012002.
  55. Equivariant transformer networks. In Chaudhuri, K. and Salakhutdinov, R. (eds.), Proceedings of the 36th International Conference on Machine Learning, volume 97 of Proceedings of Machine Learning Research, pp.  6086–6095. PMLR, 09–15 Jun 2019. URL https://proceedings.mlr.press/v97/tai19a.html.
  56. Mental rotation and orientation-dependence in shape recognition. Cognitive Psychology, 21(2):233–282, 1989. ISSN 0010-0285. doi: https://doi.org/10.1016/0010-0285(89)90009-1. URL https://www.sciencedirect.com/science/article/pii/0010028589900091.
  57. Tensor field networks: Rotation-and translation-equivariant neural networks for 3d point clouds. arXiv preprint arXiv:1802.08219, 2018.
  58. Kendall shape-VAE : Learning shapes in a generative framework. In NeurIPS 2022 Workshop on Symmetry and Geometry in Neural Representations, 2022. URL https://openreview.net/forum?id=nzh4N6kdl2G.
  59. Mdp homomorphic networks: Group symmetries in reinforcement learning. Advances in Neural Information Processing Systems, 33:4199–4210, 2020.
  60. Villani, C. Optimal transport: old and new, volume 338. Springer, 2009.
  61. Scalars are universal: Equivariant machine learning, structured like classical physics. In Beygelzimer, A., Dauphin, Y., Liang, P., and Vaughan, J. W. (eds.), Advances in Neural Information Processing Systems, 2021. URL https://openreview.net/forum?id=ba27-RzNaIv.
  62. Dynamic graph CNN for learning on point clouds. ACM Transactions on Graphics, 38(5):1–12, 2019.
  63. Unsupervised learning of group invariant and equivariant representations. In Oh, A. H., Agarwal, A., Belgrave, D., and Cho, K. (eds.), Advances in Neural Information Processing Systems, 2022. URL https://openreview.net/forum?id=47lpv23LDPr.
  64. 3d shapenets: A deep representation for volumetric shapes. pp.  1912–1920, 2015.
  65. Yarotsky, D. Universal approximations of invariant maps by neural networks. Constructive Approximation, 55(1):407–474, 2022.
  66. A rotation, scaling, and translation invariant pattern classification system. Pattern recognition, 26(5):687–710, 1993.
  67. Deep sets. In Guyon, I., Luxburg, U. V., Bengio, S., Wallach, H., Fergus, R., Vishwanathan, S., and Garnett, R. (eds.), Advances in Neural Information Processing Systems 30, pp.  3391–3401. Curran Associates, Inc., 2017a. URL http://papers.nips.cc/paper/6931-deep-sets.pdf.
  68. Deep sets. pp.  3391–3401, 2017b.
  69. Multiset-equivariant set prediction with approximate implicit differentiation. In International Conference on Learning Representations, 2022. URL https://openreview.net/forum?id=5K7RRqZEjoS.
  70. Rotation invariant convolutions for 3d point clouds deep learning. In IEEE International Conference on 3D Vision, pp.  204–213, 2019a.
  71. Shellnet: Efficient point cloud convolutional neural networks using concentric shells statistics. pp.  1607–1616, 2019b.
  72. Global context aware convolutions for 3d point cloud understanding. arXiv preprint arXiv:2008.02986, 2020.
Citations (47)
View on Semantic Scholar

Summary

We haven't generated a summary for this paper yet.

Ai Sparkles Streamline Icon: https://streamlinehq.com Summarize Now
Lightbulb Streamline Icon: https://streamlinehq.com

Continue Learning

We haven't generated follow-up questions for this paper yet.

Ai Sparkles Streamline Icon: https://streamlinehq.com Generate Now
List To Do Tasks Checklist Streamline Icon: https://streamlinehq.com

Collections

Sign up for free to add this paper to one or more collections.

User Circle Single Streamline Icon: https://streamlinehq.com Sign Up
X Twitter Logo Streamline Icon: https://streamlinehq.com

Tweets

This paper has been mentioned in 2 tweets and received 14 likes.

Upgrade to Pro to view all of the tweets about this paper:

Start a free 7-day Pro trial