Abstract
Assignment flows comprise basic dynamical systems for modeling data labeling and related machine learning tasks in supervised and unsupervised scenarios. They provide adaptive time-variant extensions of established discrete graphical models and a basis for the design and better mathematical understanding of hierarchical networks, using methods from information (differential) geometry, geometric numerical integration, statistical inference, optimal transport and control. This chapter introduces the framework by means of the image labeling problem and outlines directions of current and further research.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Notes
- 1.
This includes spatio-temporal data—like e.g. videos—observed at points \((t_{i},x_{i}) \in [0,T] \times \Omega \subset \mathbb {R} \times \mathbb {R}^{d}\) in time and space.
- 2.
The sizes of the components \(D_{i}^{j},\, j \in \mathcal {J}\) relative to each other only matter.
- 3.
Here we overload the symbol Ω which denotes the Euclidean domain covered by the graph \(\mathcal {G}\), as mentioned after Eq. (8.2). Due to the subscripts Ωi and the context, there should be no danger of confusion.
- 4.
Not to be confused with labels \(\mathcal {F}_{\ast }\)!
References
Amari, S.I., Nagaoka, H.: Methods of Information Geometry. American Mathematical Society/Oxford University Press, Providence/Oxford (2000)
Antun, V., Renna, F., Poon, C., Adcock, B., Hansen, A.C.: On instabilities of deep learning in image reconstruction: does AI come at a cost? (2019). arXiv preprint arXiv:abs/1902.05300
Åström, F., Petra, S., Schmitzer, B., Schnörr, C.: Image labeling by assignment. J. Math. Imaging Vis. 58(2), 211–238 (2017)
Ay, N., Jost, J., Lê, H.V., Schwachhöfer, L.: Information Geometry. Springer, Berlin (2017)
Barndorff-Nielsen, O.E.: Information and Exponential Families in Statistical Theory. Wiley, Chichester (1978)
Basseville, M.: Divergence measures for statistical data processing—an annotated bibliography. Signal Proc. 93(4), 621–633 (2013)
Bergmann, R., Tenbrinck, D.: A graph framework for manifold-valued data. SIAM J. Imaging Sci. 11(1), 325–360 (2018)
Berman, A., Shaked-Monderer, N.: Completely Positive Matrices. World Scientific, Singapore (2003)
Boykov, Y., Veksler, O., Zabih, R.: Fast approximate energy minimization via graph cuts. IEEE Trans. Pattern Anal. Mach. Intell. 23(11), 1222–1239 (2001)
Calin, O., Udriste, C.: Geometric Modeling in Probability and Statistics. Springer, Berlin (2014)
Chan, T.F., Esedoglu, S., Nikolova, M.: Algorithms for finding global minimizers of image segmentation and denoising models. SIAM J. Appl. Math. 66(5), 1632–1648 (2006)
Cichocki, A., Zdunek, A., Phan, A.H., Amari, S.I.: Nonnegative Matrix and Tensor Factorizations. Wiley, London (2009)
Cortes, C., Vapnik, V.: Support-vector networks. Mach. Learn. 20, 273–297 (1995)
Cover, T.M., Thomas, J.A.: Elements of Information Theory, 2nd edn. Wiley, London (2006)
Elad, M.: Deep, deep trouble: deep learning’s impact on image processing, mathematics, and humanity. SIAM News (2017)
Gary, R.M., Neuhoff, D.L.: Quantization. IEEE Trans. Inform. Theory 44(6), 2325–2383 (1998)
Geman, S., Geman, D.: Stochastic relaxation, Gibbs distributions, and the Bayesian restoration of images. IEEE Trans. Pattern Anal. Mach. Intell. 6(6), 721–741 (1984)
Graf, S., Luschgy, H.: Foundations of Quantization for Probability Distributions. Lecture Notes in Mathematics, vol. 1730. Springer, Berlin (2000)
Hairer, E., Lubich, C., Wanner, G.: Geometric Numerical Integration. Springer, Berlin (2006)
Har-Peled, S.: Geometric Approximation Algorithms. AMS, Providence (2011)
Hofbauer, J., Siegmund, K.: Evolutionary game dynamics. Bull. Am. Math. Soc. 40(4), 479–519 (2003)
Hühnerbein, R., Savarino, F., Åström, F., Schnörr, C.: Image labeling based on graphical models using Wasserstein messages and geometric assignment. SIAM J. Imaging Sci. 11(2), 1317–1362 (2018)
Hühnerbein, R., Savarino, F., Petra, S., Schnörr, C.: Learning adaptive regularization for image labeling using geometric assignment. In: Proceedings of the International Conference on Scale Space and Variational Methods in Computer Vision (SSVM). Springer, Berlin (2019)
Hummel, R.A., Zucker, S.W.: On the foundations of the relaxation labeling processes. IEEE Trans. Pattern Anal. Mach. Intell. 5(3), 267–287 (1983)
Idel, M.: A Review of Matrix scaling and Sinkhorn’s normal form for matrices and positive maps (2016). arXiv preprint arXiv:abs/1609.06349
Iserles, A., Munthe-Kaas, H.Z., Nørsett, S.P., Zanna, A.: Lie-group methods. Acta Numer. 14, 1–148 (2005)
Jost, J.: Riemannian Geometry and Geometric Analysis, 7th edn. Springer, Berlin (2017)
Kappes, J., Andres, B., Hamprecht, F., Schnörr, C., Nowozin, S., Batra, D., Kim, S., Kausler, B., Kröger, T., Lellmann, J., Komodakis, N., Savchynskyy, B., Rother, C.: A comparative study of modern inference techniques for structured discrete energy minimization problems. Int. J. Comput. Vis. 115(2), 155–184 (2015)
Kleinberg, J., Tardos, E.: Approximation algorithms for classification problems with pairwise relationships: metric labeling and Markov random fields. J. ACM 49(5), 616–639 (2002)
Koller, D., Friedman, N.: Probabilistic Graphical Models: Principles and Techniques. MIT Press, Cambridge (2009)
Krizhevsky, A., Sutskever, I., Hinton, G.E.: ImageNet classification with deep convolutional neural networks. In: Proceedings of the 25th International Conference on Neural Information Processing Systems (NIPS), pp. 1097–1105. ACM, New York (2012)
Lauritzen, S.L.: Chapter 4: statistical manifolds. In: Gupta, S.S., Amari, S.I., Barndorff-Nielsen, O.E., Kass, R.E., Lauritzen, S.L., Rao, C.R. (eds.) Differential Geometry in Statistical Inference, pp. 163–216. Institute of Mathematical Statistics, Hayward (1987)
Lauritzen, S.L.: Graphical Models. Clarendon Press, Oxford (1996)
Lee, J.M.: Introduction to Smooth Manifolds. Springer, Berlin (2013)
Lellmann, J., Schnörr, C.: Continuous multiclass labeling approaches and algorithms. SIAM J. Imag. Sci. 4(4), 1049–1096 (2011)
Mézard, M., Montanari, A.: Information, Physics, and Computation. Oxford University Press, Oxford (2009)
Munthe-Kaas, H.: High order Runge-Kutta methods on manifolds. Appl. Numer. Math. 29(1), 115–127 (1999)
Pavan, M., Pelillo, M.: Dominant sets and pairwise clustering. IEEE Trans. Pattern Anal. Mach. Intell. 29(1), 167–172 (2007)
Pelillo, M.: The dynamics of nonlinear relaxation labeling processes. J. Math. Imaging Vision 7, 309–323 (1997)
Peyré, G., Cuturi, M.: Computational Optimal Transport. CNRS, Paris (2018)
Phillips, J.: Coresets and sketches. In: Handbook of Discrete and Computational Geometry, chapter 48. CRC Press, Boca Raton (2016)
Povh, J., Rendl, F.: A copositive programming approach to graph partitioning. SIAM J. Optim. 18(1), 223–241 (2007)
Rosenfeld, A., Hummel, R.A., Zucker, S.W.: Scene labeling by relaxation operations. IEEE Trans. Syst. Man Cybern. 6, 420–433 (1976)
Ross, I.: A roadmap for optimal control: the right way to commute. Ann. N.Y. Acad. Sci. 1065(1), 210–231 (2006)
Rudin, L., Osher, S., Fatemi, E.: Nonlinear total variation based noise removal algorithms. Physica D 60, 259–268 (1992)
Rumelhart, D.E., McClelland, J.L.: Parallel Distributed Processing: Explorations in the Microstructure of Cognition: Foundations. MIT Press, Boca Raton (1986)
Sandholm, W.H.: Population Games and Evolutionary Dynamics. MIT Press, Boca Raton (2010)
Sanz-Serna, J.: Symplectic Runge–Kutta schemes for adjoint equations, automatic differentiation, optimal control, and more. SIAM Rev. 58(1), 3–33 (2016)
Savarino, F., Schnörr, C.: A variational perspective on the assignment flow. In: Proceedings of the International Conference on Scale Space and Variational Methods in Computer Vision (SSVM). Springer, Berlin (2019)
Savarino, F., Schnörr, C.: Continuous-domain assignment flows. Heidelberg University, October (2019). Preprint, submitted for publication
Wainwright, M.J., Jordan, M.I.: Graphical models, exponential families, and variational inference. Found. Trends Mach. Learn. 1(1–2), 1–305 (2008)
Wasserman, L.: All of Nonparametric Statistics. Springer, Berlin (2006)
Weinmann, A., Demaret, L., Storath, M.: Total variation regularization for manifold-valued data. SIAM J. Imag. Sci. 7(4), 2226–2257 (2014)
Werner, T.: A linear programming approach to max-sum problem: a review. IEEE Trans. Pattern Anal. Mach. Intell. 29(7), 1165–1179 (2007)
Yedidia, J.S., Freeman, W.T., Weiss, Y.: Constructing free-energy approximations and generalized belief propagation algorithms. IEEE Trans. Inform. Theory 51(7), 2282–2312 (2005)
Zeilmann, A., Savarino, F., Petra, S., Schnörr, C.: Geometric numerical integration of the assignment flow. Inverse Problems, https://doi.org/10.1088/1361-6420/ab2772 (2019, in press)
Zern, A., Zisler, M., Åström, F., Petra, S., Schnörr, C.: Unsupervised label learning on manifolds by spatially regularized geometric assignment. In: Proceedings of German Conference on Pattern Recognition (GCPR). Springer, Berlin (2018)
Zern, A., Zisler, M., Petra, S., Schnörr, C.: Unsupervised assignment flow: label learning on feature manifolds by spatially regularized geometric assignment (2019). arXiv preprint arXiv:abs/1904.10863
Zisler, M., Zern, A., Petra, S., Schnörr, C.: Unsupervised labeling by geometric and spatially regularized self-assignment. In: Proceedings of the Scale Space and Variational Methods in Computer Vision (SSVM). Springer, Berlin (2019)
Zisler, M., Zern, A., Petra, S., Schnörr, C.: Self-assignment flows for unsupervised data labeling on graphs. Heidelberg University, October (2019). Preprint, submitted for publication
Acknowledgements
I thank my students Ruben Hühnerbein, Fabrizio Savarino, Alexander Zeilmann, Artjom Zern, Matthias Zisler and my colleague Stefania Petra for many discussions and collaboration. We gratefully acknowledge support by the German Science Foundation (DFG), grant GRK 1653.
This work has also been stimulated by the recently established Heidelberg STRUCTURES Excellence Cluster, funded by the DFG under Germany’s Excellence Strategy EXC-2181/1 - 390900948.
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2020 Springer Nature Switzerland AG
About this chapter
Cite this chapter
Schnörr, C. (2020). Assignment Flows. In: Grohs, P., Holler, M., Weinmann, A. (eds) Handbook of Variational Methods for Nonlinear Geometric Data. Springer, Cham. https://doi.org/10.1007/978-3-030-31351-7_8
Download citation
DOI: https://doi.org/10.1007/978-3-030-31351-7_8
Published:
Publisher Name: Springer, Cham
Print ISBN: 978-3-030-31350-0
Online ISBN: 978-3-030-31351-7
eBook Packages: Mathematics and StatisticsMathematics and Statistics (R0)