Deep learning as optimal control problems: Models and numerical methods

Previous Article
Algebraic structure of aromatic B-series
JCD Home
This Issue
Next Article
Efficient time integration methods for Gross-Pitaevskii equations with rotation term

December 2019, 6(2): 171-198. doi: 10.3934/jcd.2019009

Deep learning as optimal control problems: Models and numerical methods

Martin Benning ^1, , Elena Celledoni ^2, , Matthias J. Ehrhardt ^3, , Brynjulf Owren ^1, and Carola-Bibiane Schönlieb ^4,,

1.	School of Mathematical Sciences, Queen Mary University of London, London E1 4NS, UK
2.	Department of Mathematical Sciences, NTNU, 7491 Trondheim, Norway
3.	Institute for Mathematical Innovation, University of Bath, Bath BA2 7JU, UK
4.	Department of Applied Mathematics and Theoretical Physics, University of Cambridge, Cambridge CB3 0WA, UK

^* Corresponding author: Carola-Bibiane Schönlieb

Received April 2019 Revised September 2019 Published November 2019

We consider recent work of [18] and [9], where deep learning neural networks have been interpreted as discretisations of an optimal control problem subject to an ordinary differential equation constraint. We review the first order conditions for optimality, and the conditions ensuring optimality after discretisation. This leads to a class of algorithms for solving the discrete optimal control problem which guarantee that the corresponding discrete necessary conditions for optimality are fulfilled. The differential equation setting lends itself to learning additional parameters such as the time discretisation. We explore this extension alongside natural constraints (e.g. time steps lie in a simplex). We compare these deep learning algorithms numerically in terms of induced flow and generalisation ability.

Keywords: Deep learning, optimal control, Runge–Kutta methods, Hamiltonian boundary value problems.

Mathematics Subject Classification: Primary: 65L10, 65K10, 68Q32; Secondary: 49J15.

Citation: Martin Benning, Elena Celledoni, Matthias J. Ehrhardt, Brynjulf Owren, Carola-Bibiane Schönlieb. Deep learning as optimal control problems: Models and numerical methods. Journal of Computational Dynamics, 2019, 6 (2) : 171-198. doi: 10.3934/jcd.2019009

References:

[1]	H. D. Abarbanel, P. J. Rozdeba and S. Shirman, Machine learning: Deepest learning as statistical data assimilation problems, Neural Comput., 30 (2018), 2025-2055. doi: 10.1162/neco_a_01094. Google Scholar
[2]	A. A. Agrachev and Y. Sachkov, Control Theory from the Geometric Viewpoint, Encyclopaedia of Mathematical Sciences, 87, Springer-Verlag, Berlin, 2004. doi: 10.1007/978-3-662-06404-7. Google Scholar
[3]	M. Benning and M. Burger, Modern regularization methods for inverse problems, Acta Numer., 27 (2018), 1-111. doi: 10.1017/S0962492918000016. Google Scholar
[4]	C. M. Bishop, Pattern Recognition and Machine Learning, Information Science and Statistics, Springer, New York, 2006. Google Scholar
[5]	S. Blanes and F. Casas, On the necessity of negative coefficients for operator splitting schemes of order higher than two, Appl. Numer. Math., 54 (2005), 23-37. doi: 10.1016/j.apnum.2004.10.005. Google Scholar
[6]	R. Bucy, Two-point boundary value problems of linear Hamiltonian systems, SIAM J. Appl. Math., 15 (1967), 1385-1389. doi: 10.1137/0115121. Google Scholar
[7]	M. Burger, G. Gilboa, S. Osher and J. Xu et al., Nonlinear inverse scale space methods, Commun. Math. Sci., 4 (2006), 179-212. doi: 10.4310/CMS.2006.v4.n1.a7. Google Scholar
[8]	M. Burger, S. Osher, J. Xu and G. Gilboa, Nonlinear inverse scale space methods for image restoration, in International Workshop on Variational, Geometric, and Level Set Methods in Computer Vision, Lecture Notes in Computer Science, 3752, Springer-Verlag, Berlin, 2005, 25–36. doi: 10.1007/11567646_3. Google Scholar
[9]	B. Chang, L. Meng, E. Haber, L. Ruthotto, D. Begert and E. Holtham, Reversible architectures for arbitrarily deep residual neural networks, 32^nd AAAI Conference on Artificial Intelligence, 2018. Google Scholar
[10]	T. Q. Chen, Y. Rubanova, J. Bettencourt and D. K. Duvenaud, Neural ordinary differential equations, in Advances in Neural Information Processing Systems, 2018, 6572–6583. Google Scholar
[11]	Y. N. Dauphin, R. Pascanu, C. Gulcehre, K. Cho, S. Ganguli and Y. Bengio, Identifying and attacking the saddle point problem in high-dimensional non-convex optimization, in Advances in Neural Information Processing Systems, 2014, 2933–2941. Google Scholar
[12]	A. L. Dontchev and W. W. Hager, The Euler approximation in state constrained optimal control, Math. Comp., 70 (2001), 173-203. doi: 10.1090/S0025-5718-00-01184-4. Google Scholar
[13]	J. Duchi, S. Shalev-Shwartz, Y. Singer and T. Chandra, Efficient projections onto the $\ell$1-ball for learning in high dimensions, in Proceedings of the 25^th International Conference on Machine Learning, 2008, 272–279. Google Scholar
[14]	W. E, A proposal on machine learning via dynamical systems, Commun. Math. Stat., 5 (2017), 1-11. doi: 10.1007/s40304-017-0103-z. Google Scholar
[15]	F. Gay-Balmaz and T. S. Ratiu, Clebsch optimal control formulation in mechanics, J. Geom. Mech., 3 (2011), 41-79. doi: 10.3934/jgm.2011.3.41. Google Scholar
[16]	A. Gholami, K. Keutzer and G. Biros, ANODE: Unconditionally accurate memory-efficient gradients for neural ODEs, preprint, arXiv: math/1902.10298. Google Scholar
[17]	I. Goodfellow, J. Shlens and C. Szegedy, Explaining and harnessing adversarial examples, preprint, arXiv: math/1412.6572. Google Scholar
[18]	E. Haber and L. Ruthotto, Stable architectures for deep neural networks, Inverse Problems, 34 (2018), 22pp. doi: 10.1088/1361-6420/aa9a90. Google Scholar
[19]	W. W. Hager, Runge-Kutta methods in optimal control and the transformed adjoint system, Numer. Math., 87 (2000), 247-282. doi: 10.1007/s002110000178. Google Scholar
[20]	E. Hairer, C. Lubich and G. Wanner, Geometric Numerical Integration: Structure-Preserving Algorithms for Ordinary Differential Equations, Springer Series in Computational Mathematics, 31, Springer-Verlag, Berlin, 2006. doi: 10.1007/3-540-30666-8. Google Scholar
[21]	K. He, X. Zhang, S. Ren and J. Sun, Identity mappings in deep residual networks, in European Conference on Computer Vision, Lecture Notes in Computer Science, 9908, Springer, Cham, 2016, 630–645. doi: 10.1007/978-3-319-46493-0_38. Google Scholar
[22]	K. He, X. Zhang, S. Ren and J. Sun, Deep residual learning for image recognition, IEEE Conference on Computer Vision and Pattern Recognition, Las Vegas, NV, 2016, 770–778. doi: 10.1109/CVPR.2016.90. Google Scholar
[23]	C. F. Higham and D. J. Higham, Deep learning: An introduction for applied mathematicians, preprint, arXiv: math/1801.05894. Google Scholar
[24]	M. Igami, Artificial intelligence as structural estimation: Economic interpretations of Deep Blue, Bonanza, and AlphaGo, preprint, arXiv: math/1710.10967. Google Scholar
[25]	A. Kurakin, I. Goodfellow and S. Bengio, Adversarial examples in the physical world, preprint, arXiv: math/1607.02533. Google Scholar
[26]	W. Kutta, Beitrag zur näherungsweisen integration totaler differentialgleichungen, Z. Math. Phys., 46 (1901), 435-453. Google Scholar
[27]	Y. LeCun, Y. Bengio and G. Hinton, Deep learning, Nature, 521 (2015), 436-444. doi: 10.1038/nature14539. Google Scholar
[28]	Y. LeCun, L. Bottou, Y. Bengio and P. Haffner, Gradient-based learning applied to document recognition, Proceedings of the IEEE, 86, 1998, 2278–2324. doi: 10.1109/5.726791. Google Scholar
[29]	Y. LeCun, A theoretical framework for back-propagation, Proceedings of the 1988 Connectionist Models Summer School, 1, CMU, Pittsburgh, PA, 1988, 21–28. Google Scholar
[30]	Q. Li, L. Chen, C. Tai and E. Weinan, Maximum principle based algorithms for deep learning, J. Mach. Learn. Res., 18 (2017), 5998-6026. Google Scholar
[31]	Q. Li and S. Hao, An optimal control approach to deep learning and applications to discrete-weight neural networks, preprint, arXiv: math/1803.01299. doi: 10.1109/TNNLS.2017.2665555. Google Scholar
[32]	Y. Lu, A. Zhong, Q. Li and B. Dong, Beyond finite layer neural networks: Bridging deep architectures and numerical differential equations, in Proceedings of the 35^th International Conference on Machine Learning, Proceedings of Machine Learning Research, 80, Stockholm, Sweden, 2018, 3276–3285. Available at: http://proceedings.mlr.press/v80/lu18d.html. Google Scholar
[33]	S.-M. Moosavi-Dezfooli, A. Fawzi and P. Frossard, Deepfool: A simple and accurate method to fool deep neural networks, in Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2016, 2574–2582. doi: 10.1109/CVPR.2016.282. Google Scholar
[34]	D. L. Phillips, A technique for the numerical solution of certain integral equations of the first kind, J. Assoc. Comput. Mach., 9 (1962), 84-97. doi: 10.1145/321105.321114. Google Scholar
[35]	D. C. Plaut et al., Experiments on learning by back propagation, work in progress. Google Scholar
[36]	L. Pontryagin et. al, Selected Works: The Mathematical Theory of Optimal Processes, Classics of Soviet Mathematics, 4, Gordon & Breach Science Publishers, New York, 1986. Google Scholar
[37]	I. M. Ross, A roadmap for optimal control: The right way to commute, Ann. N.Y. Acad. Sci., 1065 (2005), 210-231. doi: 10.1196/annals.1370.015. Google Scholar
[38]	F. Santosa and W. W. Symes, Linear inversion of band-limited reflection seismograms, SIAM J. Sci. Statist. Comput., 7 (1986), 1307-1330. doi: 10.1137/0907087. Google Scholar
[39]	J. M. Sanz-Serna, Symplectic Runge-Kutta schemes for adjoint equations automatic differentiation, optimal control and more, SIAM Rev., 58 (2016), 3-33. doi: 10.1137/151002769. Google Scholar
[40]	O. Scherzer and C. Groetsch, Inverse scale space theory for inverse problems, in International Conference on Scale-Space Theories in Computer Vision, Lecture Notes in Computer Science, 2106, Springer, Berlin, 2001, 317–325. doi: 10.1007/3-540-47778-0_29. Google Scholar
[41]	S. Sonoda and N. Murata, Double continuum limit of deep neural networks, ICML Workshop Principled Approaches to Deep Learning, 2017. Google Scholar
[42]	E. D. Sontag, Mathematical Control Theory: Deterministic Finite-Dimensional Systems, Texts in Applied Mathematics, 6, Springer-Verlag, New York, 1998. doi: 10.1007/978-1-4612-0577-7. Google Scholar
[43]	I. Sutskever, J. Martens, G. E. Dahl and G. E. Hinton, On the importance of initialization and momentum in deep learning, Proceedings of the 30^th International Conference on Machine Learning, 28, 2013, 1139–1147. Google Scholar
[44]	C. Szegedy, W. Zaremba, I. Sutskever, J. Bruna, D. Erhan, I. J. Goodfellow and R. Fergus, Intriguing properties of neural networks, International Conference on Learning Representations, 2014. Google Scholar
[45]	M. Thorpe and Y. van Gennip, Deep limits of residual neural networks, preprint, arXiv: math/1810.11741. Google Scholar
[46]	R. Tibshirani, Regression shrinkage and selection via the lasso, J. Roy. Statist. Soc. Ser. B, 58 (1996), 267-288. doi: 10.1111/j.2517-6161.1996.tb02080.x. Google Scholar
[47]	A. N. Tikhonov, Solution of incorrectly formulated problems and the regularization method, in Dokl. Akad. Nauk., 151, 1963, 1035–1038. Google Scholar
[48]	E. Weinan, J. Han and Q. Li, A mean-field optimal control formulation of deep learning, Res. Math. Sci., 6 (2019), 41pp. doi: 10.1007/s40687-018-0172-y. Google Scholar

show all references

References:

[1]	H. D. Abarbanel, P. J. Rozdeba and S. Shirman, Machine learning: Deepest learning as statistical data assimilation problems, Neural Comput., 30 (2018), 2025-2055. doi: 10.1162/neco_a_01094. Google Scholar
[2]	A. A. Agrachev and Y. Sachkov, Control Theory from the Geometric Viewpoint, Encyclopaedia of Mathematical Sciences, 87, Springer-Verlag, Berlin, 2004. doi: 10.1007/978-3-662-06404-7. Google Scholar
[3]	M. Benning and M. Burger, Modern regularization methods for inverse problems, Acta Numer., 27 (2018), 1-111. doi: 10.1017/S0962492918000016. Google Scholar
[4]	C. M. Bishop, Pattern Recognition and Machine Learning, Information Science and Statistics, Springer, New York, 2006. Google Scholar
[5]	S. Blanes and F. Casas, On the necessity of negative coefficients for operator splitting schemes of order higher than two, Appl. Numer. Math., 54 (2005), 23-37. doi: 10.1016/j.apnum.2004.10.005. Google Scholar
[6]	R. Bucy, Two-point boundary value problems of linear Hamiltonian systems, SIAM J. Appl. Math., 15 (1967), 1385-1389. doi: 10.1137/0115121. Google Scholar
[7]	M. Burger, G. Gilboa, S. Osher and J. Xu et al., Nonlinear inverse scale space methods, Commun. Math. Sci., 4 (2006), 179-212. doi: 10.4310/CMS.2006.v4.n1.a7. Google Scholar
[8]	M. Burger, S. Osher, J. Xu and G. Gilboa, Nonlinear inverse scale space methods for image restoration, in International Workshop on Variational, Geometric, and Level Set Methods in Computer Vision, Lecture Notes in Computer Science, 3752, Springer-Verlag, Berlin, 2005, 25–36. doi: 10.1007/11567646_3. Google Scholar
[9]	B. Chang, L. Meng, E. Haber, L. Ruthotto, D. Begert and E. Holtham, Reversible architectures for arbitrarily deep residual neural networks, 32^nd AAAI Conference on Artificial Intelligence, 2018. Google Scholar
[10]	T. Q. Chen, Y. Rubanova, J. Bettencourt and D. K. Duvenaud, Neural ordinary differential equations, in Advances in Neural Information Processing Systems, 2018, 6572–6583. Google Scholar
[11]	Y. N. Dauphin, R. Pascanu, C. Gulcehre, K. Cho, S. Ganguli and Y. Bengio, Identifying and attacking the saddle point problem in high-dimensional non-convex optimization, in Advances in Neural Information Processing Systems, 2014, 2933–2941. Google Scholar
[12]	A. L. Dontchev and W. W. Hager, The Euler approximation in state constrained optimal control, Math. Comp., 70 (2001), 173-203. doi: 10.1090/S0025-5718-00-01184-4. Google Scholar
[13]	J. Duchi, S. Shalev-Shwartz, Y. Singer and T. Chandra, Efficient projections onto the $\ell$1-ball for learning in high dimensions, in Proceedings of the 25^th International Conference on Machine Learning, 2008, 272–279. Google Scholar
[14]	W. E, A proposal on machine learning via dynamical systems, Commun. Math. Stat., 5 (2017), 1-11. doi: 10.1007/s40304-017-0103-z. Google Scholar
[15]	F. Gay-Balmaz and T. S. Ratiu, Clebsch optimal control formulation in mechanics, J. Geom. Mech., 3 (2011), 41-79. doi: 10.3934/jgm.2011.3.41. Google Scholar
[16]	A. Gholami, K. Keutzer and G. Biros, ANODE: Unconditionally accurate memory-efficient gradients for neural ODEs, preprint, arXiv: math/1902.10298. Google Scholar
[17]	I. Goodfellow, J. Shlens and C. Szegedy, Explaining and harnessing adversarial examples, preprint, arXiv: math/1412.6572. Google Scholar
[18]	E. Haber and L. Ruthotto, Stable architectures for deep neural networks, Inverse Problems, 34 (2018), 22pp. doi: 10.1088/1361-6420/aa9a90. Google Scholar
[19]	W. W. Hager, Runge-Kutta methods in optimal control and the transformed adjoint system, Numer. Math., 87 (2000), 247-282. doi: 10.1007/s002110000178. Google Scholar
[20]	E. Hairer, C. Lubich and G. Wanner, Geometric Numerical Integration: Structure-Preserving Algorithms for Ordinary Differential Equations, Springer Series in Computational Mathematics, 31, Springer-Verlag, Berlin, 2006. doi: 10.1007/3-540-30666-8. Google Scholar
[21]	K. He, X. Zhang, S. Ren and J. Sun, Identity mappings in deep residual networks, in European Conference on Computer Vision, Lecture Notes in Computer Science, 9908, Springer, Cham, 2016, 630–645. doi: 10.1007/978-3-319-46493-0_38. Google Scholar
[22]	K. He, X. Zhang, S. Ren and J. Sun, Deep residual learning for image recognition, IEEE Conference on Computer Vision and Pattern Recognition, Las Vegas, NV, 2016, 770–778. doi: 10.1109/CVPR.2016.90. Google Scholar
[23]	C. F. Higham and D. J. Higham, Deep learning: An introduction for applied mathematicians, preprint, arXiv: math/1801.05894. Google Scholar
[24]	M. Igami, Artificial intelligence as structural estimation: Economic interpretations of Deep Blue, Bonanza, and AlphaGo, preprint, arXiv: math/1710.10967. Google Scholar
[25]	A. Kurakin, I. Goodfellow and S. Bengio, Adversarial examples in the physical world, preprint, arXiv: math/1607.02533. Google Scholar
[26]	W. Kutta, Beitrag zur näherungsweisen integration totaler differentialgleichungen, Z. Math. Phys., 46 (1901), 435-453. Google Scholar
[27]	Y. LeCun, Y. Bengio and G. Hinton, Deep learning, Nature, 521 (2015), 436-444. doi: 10.1038/nature14539. Google Scholar
[28]	Y. LeCun, L. Bottou, Y. Bengio and P. Haffner, Gradient-based learning applied to document recognition, Proceedings of the IEEE, 86, 1998, 2278–2324. doi: 10.1109/5.726791. Google Scholar
[29]	Y. LeCun, A theoretical framework for back-propagation, Proceedings of the 1988 Connectionist Models Summer School, 1, CMU, Pittsburgh, PA, 1988, 21–28. Google Scholar
[30]	Q. Li, L. Chen, C. Tai and E. Weinan, Maximum principle based algorithms for deep learning, J. Mach. Learn. Res., 18 (2017), 5998-6026. Google Scholar
[31]	Q. Li and S. Hao, An optimal control approach to deep learning and applications to discrete-weight neural networks, preprint, arXiv: math/1803.01299. doi: 10.1109/TNNLS.2017.2665555. Google Scholar
[32]	Y. Lu, A. Zhong, Q. Li and B. Dong, Beyond finite layer neural networks: Bridging deep architectures and numerical differential equations, in Proceedings of the 35^th International Conference on Machine Learning, Proceedings of Machine Learning Research, 80, Stockholm, Sweden, 2018, 3276–3285. Available at: http://proceedings.mlr.press/v80/lu18d.html. Google Scholar
[33]	S.-M. Moosavi-Dezfooli, A. Fawzi and P. Frossard, Deepfool: A simple and accurate method to fool deep neural networks, in Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2016, 2574–2582. doi: 10.1109/CVPR.2016.282. Google Scholar
[34]	D. L. Phillips, A technique for the numerical solution of certain integral equations of the first kind, J. Assoc. Comput. Mach., 9 (1962), 84-97. doi: 10.1145/321105.321114. Google Scholar
[35]	D. C. Plaut et al., Experiments on learning by back propagation, work in progress. Google Scholar
[36]	L. Pontryagin et. al, Selected Works: The Mathematical Theory of Optimal Processes, Classics of Soviet Mathematics, 4, Gordon & Breach Science Publishers, New York, 1986. Google Scholar
[37]	I. M. Ross, A roadmap for optimal control: The right way to commute, Ann. N.Y. Acad. Sci., 1065 (2005), 210-231. doi: 10.1196/annals.1370.015. Google Scholar
[38]	F. Santosa and W. W. Symes, Linear inversion of band-limited reflection seismograms, SIAM J. Sci. Statist. Comput., 7 (1986), 1307-1330. doi: 10.1137/0907087. Google Scholar
[39]	J. M. Sanz-Serna, Symplectic Runge-Kutta schemes for adjoint equations automatic differentiation, optimal control and more, SIAM Rev., 58 (2016), 3-33. doi: 10.1137/151002769. Google Scholar
[40]	O. Scherzer and C. Groetsch, Inverse scale space theory for inverse problems, in International Conference on Scale-Space Theories in Computer Vision, Lecture Notes in Computer Science, 2106, Springer, Berlin, 2001, 317–325. doi: 10.1007/3-540-47778-0_29. Google Scholar
[41]	S. Sonoda and N. Murata, Double continuum limit of deep neural networks, ICML Workshop Principled Approaches to Deep Learning, 2017. Google Scholar
[42]	E. D. Sontag, Mathematical Control Theory: Deterministic Finite-Dimensional Systems, Texts in Applied Mathematics, 6, Springer-Verlag, New York, 1998. doi: 10.1007/978-1-4612-0577-7. Google Scholar
[43]	I. Sutskever, J. Martens, G. E. Dahl and G. E. Hinton, On the importance of initialization and momentum in deep learning, Proceedings of the 30^th International Conference on Machine Learning, 28, 2013, 1139–1147. Google Scholar
[44]	C. Szegedy, W. Zaremba, I. Sutskever, J. Bruna, D. Erhan, I. J. Goodfellow and R. Fergus, Intriguing properties of neural networks, International Conference on Learning Representations, 2014. Google Scholar
[45]	M. Thorpe and Y. van Gennip, Deep limits of residual neural networks, preprint, arXiv: math/1810.11741. Google Scholar
[46]	R. Tibshirani, Regression shrinkage and selection via the lasso, J. Roy. Statist. Soc. Ser. B, 58 (1996), 267-288. doi: 10.1111/j.2517-6161.1996.tb02080.x. Google Scholar
[47]	A. N. Tikhonov, Solution of incorrectly formulated problems and the regularization method, in Dokl. Akad. Nauk., 151, 1963, 1035–1038. Google Scholar
[48]	E. Weinan, J. Han and Q. Li, A mean-field optimal control formulation of deep learning, Res. Math. Sci., 6 (2019), 41pp. doi: 10.1007/s40687-018-0172-y. Google Scholar

Figure 1. The four data sets used in the numerical study

Download full-size image

Download as PowerPoint slide

Figure 2. Learned transformation and classifier for data set donut1d (top) and squares (bottom)

Download full-size image

Download as PowerPoint slide

Figure 3. Snap shots of transformation of features for data set spiral

Download full-size image

Download as PowerPoint slide

Figure 4. Learned transformation with fixed classifier for data set donut1d (top) and spiral (bottom)

Download full-size image

Download as PowerPoint slide

Figure 5. Robustness on random initialisation for transformed data donut2d and linear classifier for two different initialisations

Download full-size image

Download as PowerPoint slide

Figure 6. Function values over the course of the gradient descent iterations for data sets donut1d, donut2d, spiral, squares (left to right and top to bottom). The solid line represents training and the dashed line test data

Download full-size image

Download as PowerPoint slide

Figure 7. Classification accuracy over the course of the gradient descent iterations for data setsdonut1d, donut2d, spiral, squares (left to right and top to bottom). The solid line represents training and the dashed line test data

Download full-size image

Download as PowerPoint slide

Figure 8. Estimated time steps by ResNet/Euler, ODENet and ODENet+simplex for for data sets donut1d, donut2d, spiral, squares (left to right and top to bottom). ODENet+simplex consistently picks two to three time steps and set the rest to zero

Download full-size image

Download as PowerPoint slide

Figure 9. Learned prediction and transformation for different Runge-Kutta methods and data sets spiral (top), donut2d (centre) and squares (bottom). All results are for 15 layers

Download full-size image

Download as PowerPoint slide

Figure 10. Snap shots of the transition from initial to final state through the network with the data set spiral

Download full-size image

Download as PowerPoint slide

Figure 11. Snap shots of the transition from initial to final state through the network with the data set donut2d

Download full-size image

Download as PowerPoint slide

Figure 12. Snap shots of the transition from initial to final state through the network with the data set squares

Download full-size image

Download as PowerPoint slide

Figure 13. Function values over the course of the gradient descent iterations for data sets donut1d, donut2d, spiral, squares (left to right and top to bottom)

Download full-size image

Download as PowerPoint slide

Figure 14. Accuracy (left) and time steps (right) for MNIST100 dataset [28]

Download full-size image

Download as PowerPoint slide

Figure 15. Features of testing examples from MNIST100 dataset [28] and transformed features by four networks under comparison: Net, ResNet, ODENet, ODENet+Simplex (from top to bottom). All networks have 20 layers

Download full-size image

Download as PowerPoint slide

Table 1. Four explicit Runge–Kutta methods: ResNet/Euler, Improved Euler, Kutta(3) and Kutta(4)

Download as excel

[1]	Sihong Shao, Huazhong Tang. Higher-order accurate Runge-Kutta discontinuous Galerkin methods for a nonlinear Dirac model. Discrete & Continuous Dynamical Systems - B, 2006, 6 (3) : 623-640. doi: 10.3934/dcdsb.2006.6.623
[2]	Antonia Katzouraki, Tania Stathaki. Intelligent traffic control on internet-like topologies - integration of graph principles to the classic Runge--Kutta method. Conference Publications, 2009, 2009 (Special) : 404-415. doi: 10.3934/proc.2009.2009.404
[3]	Wenjuan Zhai, Bingzhen Chen. A fourth order implicit symmetric and symplectic exponentially fitted Runge-Kutta-Nyström method for solving oscillatory problems. Numerical Algebra, Control & Optimization, 2019, 9 (1) : 71-84. doi: 10.3934/naco.2019006
[4]	K. Q. Lan, G. C. Yang. Optimal constants for two point boundary value problems. Conference Publications, 2007, 2007 (Special) : 624-633. doi: 10.3934/proc.2007.2007.624
[5]	Chunjuan Hou, Yanping Chen, Zuliang Lu. Superconvergence property of finite element methods for parabolic optimal control problems. Journal of Industrial & Management Optimization, 2011, 7 (4) : 927-945. doi: 10.3934/jimo.2011.7.927
[6]	Z. Foroozandeh, Maria do rosário de Pinho, M. Shamsi. On numerical methods for singular optimal control problems: An application to an AUV problem. Discrete & Continuous Dynamical Systems - B, 2019, 24 (5) : 2219-2235. doi: 10.3934/dcdsb.2019092
[7]	Alexander Mielke. Weak-convergence methods for Hamiltonian multiscale problems. Discrete & Continuous Dynamical Systems - A, 2008, 20 (1) : 53-79. doi: 10.3934/dcds.2008.20.53
[8]	Monica Motta, Caterina Sartori. Uniqueness results for boundary value problems arising from finite fuel and other singular and unbounded stochastic control problems. Discrete & Continuous Dynamical Systems - A, 2008, 21 (2) : 513-535. doi: 10.3934/dcds.2008.21.513
[9]	Ugur G. Abdulla, Evan Cosgrove, Jonathan Goldfarb. On the Frechet differentiability in optimal control of coefficients in parabolic free boundary problems. Evolution Equations & Control Theory, 2017, 6 (3) : 319-344. doi: 10.3934/eect.2017017
[10]	Da Xu. Numerical solutions of viscoelastic bending wave equations with two term time kernels by Runge-Kutta convolution quadrature. Discrete & Continuous Dynamical Systems - B, 2017, 22 (6) : 2389-2416. doi: 10.3934/dcdsb.2017122
[11]	Pierluigi Colli, Gianni Gilardi, Jürgen Sprekels. Deep quench approximation and optimal control of general Cahn–Hilliard systems with fractional operators and double obstacle potentials. Discrete & Continuous Dynamical Systems - S, 2018, 0 (0) : 0-0. doi: 10.3934/dcdss.2020213
[12]	Colin J. Cotter, Darryl D. Holm. Geodesic boundary value problems with symmetry. Journal of Geometric Mechanics, 2010, 2 (1) : 51-68. doi: 10.3934/jgm.2010.2.51
[13]	Olha P. Kupenko, Rosanna Manzo. On optimal controls in coefficients for ill-posed non-Linear elliptic Dirichlet boundary value problems. Discrete & Continuous Dynamical Systems - B, 2018, 23 (4) : 1363-1393. doi: 10.3934/dcdsb.2018155
[14]	Orazio Arena. On some boundary control problems. Discrete & Continuous Dynamical Systems - S, 2016, 9 (3) : 613-618. doi: 10.3934/dcdss.2016015
[15]	Xiaoxi Li, Marc Quincampoix, Jérôme Renault. Limit value for optimal control with general means. Discrete & Continuous Dynamical Systems - A, 2016, 36 (4) : 2113-2132. doi: 10.3934/dcds.2016.36.2113
[16]	Laurence Halpern, Jeffrey Rauch. Hyperbolic boundary value problems with trihedral corners. Discrete & Continuous Dynamical Systems - A, 2016, 36 (8) : 4403-4450. doi: 10.3934/dcds.2016.36.4403
[17]	Felix Sadyrbaev. Nonlinear boundary value problems of the calculus of variations. Conference Publications, 2003, 2003 (Special) : 760-770. doi: 10.3934/proc.2003.2003.760
[18]	J. R. L. Webb. Uniqueness of the principal eigenvalue in nonlocal boundary value problems. Discrete & Continuous Dynamical Systems - S, 2008, 1 (1) : 177-186. doi: 10.3934/dcdss.2008.1.177
[19]	Sergei Avdonin, Fritz Gesztesy, Konstantin A. Makarov. Spectral estimation and inverse initial boundary value problems. Inverse Problems & Imaging, 2010, 4 (1) : 1-9. doi: 10.3934/ipi.2010.4.1
[20]	G. Infante. Positive solutions of nonlocal boundary value problems with singularities. Conference Publications, 2009, 2009 (Special) : 377-384. doi: 10.3934/proc.2009.2009.377

Impact Factor:

Tools

Metrics

Other articles
by authors

on AIMS
Martin Benning Elena Celledoni Matthias J. Ehrhardt Brynjulf Owren Carola-Bibiane Schönlieb
on Google Scholar
Martin Benning Elena Celledoni Matthias J. Ehrhardt Brynjulf Owren Carola-Bibiane Schönlieb

Impact Factor:

Tools

Article outline

Figures and Tables

Impact Factor:

Tools

Figures and Tables

Impact Factor:

Tools

Metrics

Other articles
by authors

on AIMS
Martin Benning Elena Celledoni Matthias J. Ehrhardt Brynjulf Owren Carola-Bibiane Schönlieb
on Google Scholar
Martin Benning Elena Celledoni Matthias J. Ehrhardt Brynjulf Owren Carola-Bibiane Schönlieb

[Back to Top]