Papers
Topics
Authors
Recent
Gemini 2.5 Flash
Gemini 2.5 Flash
166 tokens/sec
GPT-4o
7 tokens/sec
Gemini 2.5 Pro Pro
42 tokens/sec
o3 Pro
4 tokens/sec
GPT-4.1 Pro
38 tokens/sec
DeepSeek R1 via Azure Pro
28 tokens/sec
2000 character limit reached

A kernel-based method for Schrödinger bridges (2310.14522v2)

Published 23 Oct 2023 in math.OC, cs.NA, and math.NA

Abstract: We characterize the Schr\"odinger bridge problems by a family of Mckean-Vlasov stochastic control problems with no terminal time distribution constraint. In doing so, we use the theory of Hilbert space embeddings of probability measures and then describe the constraint as penalty terms defined by the maximum mean discrepancy in the control problems. A sequence of the probability laws of the state processes resulting from $\epsilon$-optimal controls converges to a unique solution of the Schr\"odinger's problem under mild conditions on given initial and terminal time distributions and an underlying diffusion process. We propose a neural SDE based deep learning algorithm for the Mckean-Vlasov stochastic control problems. Several numerical experiments validate our methods.

Definition Search Book Streamline Icon: https://streamlinehq.com
References (45)
  1. B. D. Anderson. Reverse-time diffusion equation models. Stochastic Processes and their Applications, 12:313–326, 1982.
  2. D. G. Aronson. Bounds for the fundamental solution of a parabolic equation. Bull. Amer. Math. Soc., 73:890–896, 1967.
  3. S. Bernstein. Sur les liaisons entre les grandeurs aléatoires. In Proc. Int. Cong. of Math., volume 1, pages 288–309, 1932.
  4. A. Beurling. An automorphism of product measures. Ann. Math., 72:189–200, 1960.
  5. P. Billingsley. Probability and measure. John Wiley & Sons, New York, 3rd edition, 1995.
  6. P. Billingsley. Convergence of Probability Measures. John Wiley & Sons, New York, 2nd edition, 1999.
  7. C. M. Bishop. Pattern recognition and machine learning. Springer, New York, 2006.
  8. Entropic and displacement interpolation: a computational approach using the hilbert metric. SIAM J. Appl. Math., 76:2375–2396, 2016.
  9. Stochastic control liaisons: Richard Sinkhorn meets Gaspard Monge on a Schrödinger bridge. SIAM Rev., 63:249–313, 2021.
  10. E. Schrödinger’s 1931 paper “On the Reversal of the Laws of Nature” [“Über die Umkehrung der Naturgesetze”, Sitzungsberichte der preussischen Akademie der Wissenschaften, physikalisch-mathematische Klasse, 8 N9 144–153]. Eur. Phys. J. H, 46:1–29, 2021.
  11. P. Dai Pra. A stochastic control approach to reciprocal diffusion processes. Appl. Math. Optim., 23:313–329, 1991.
  12. Diffusion Schrödinger bridge with applications to score-based generative modeling. Advances in Neural Information Processing Systems, 34:17695–17709, 2021.
  13. A. Dembo and O. Zeitouni. Large deviations techniques and applications. Springer, 1998.
  14. Controlled Markov processes and viscosity solutions. Springer-Verlag, New York, 2nd edition, 2006.
  15. H. Föllmer. An entropy approach to the time reversal of diffusion processes. In Stochastic Differential Systems Filtering and Control, pages 156–163. Springer, 1985.
  16. H. Föllmer. Time reversal on Wiener space. In Stochastic processes-mathematics and physics, pages 119–129. Springer, 1986.
  17. H. Föllmer. Random fields and diffusion processes. In École d’Été de Probabilités de Saint-Flour XV–XVII, 1985–87, pages 101–203. Springer, 1988.
  18. Deep learning. MIT press, 2016.
  19. A kernel method for the two-sample-problem. Advances in neural information processing systems, 19, 2006.
  20. Denoising diffusion probabilistic models. Advances in Neural Information Processing Systems, 33:6840–6851, 2020.
  21. Schrödinger-Föllmer sampler: sampling without ergodicity. arXiv:2106.10880[stat.CO].
  22. K. Itô. An introduction to probability theory. Cambridge University Press, Cambridge, 1984.
  23. B. Jamison. Reciprocal processes. Z. Wahrscheinlichkeitstheorie verw. Gebiete, 30:65–86, 1974.
  24. B. Jamison. The Markov processes of Schrödinger. Z. Wahrscheinlichkeitstheorie verw. Gebiete, 32:323–331, 1975.
  25. I. Karatzas and S. E. Shreve. Brownian motion and stochastic calculus. Springer-Verlag, New York, 1991.
  26. Efficient and accurate gradients for neural SDEs. Advances in Neural Information Processing Systems, 34:18747–18761, 2021.
  27. Neural SDEs as infinite-dimensional GANs. In International conference on machine learning, pages 5453–5463. PMLR, 2021.
  28. C. Léonard. Revisiting Fortet’s proof of existence of a solution to the Schrödinger system. arXiv:1904.13211[Math.PR].
  29. C. Léonard. A survey of the Schrödinger problem and some of its connections with optimal transport. Discrete Contin. Dyn. Syst., 34:1533–1574, 2013.
  30. X. Li. torchsde, 2020. https://github.com/google-research/torchsde.
  31. R. Liptser and A. N. Shiryaev. Statistics of random processes: I. General theory. Springer, Berlin, 2nd rev. and exp. edition, 2001.
  32. D. G. Luenberger. Optimization by vector space methods. John Wiley & Sons, New York, 1997.
  33. T. Mikami. Monge’s problem with a quadratic cost by the zero-noise limit of hℎhitalic_h-path processes. Probab. Theory Related Fields, 129:245–260, 2004.
  34. T. Mikami and M. Thieullen. Duality theorem for the stochastic optimal control problem. Stochastic Process. Appl., 116(12):1815–1835, 2006.
  35. M. Nagasawa. Transformations of diffusion and Schrödinger processes. Probab. Theory Related Fields, 82:109–136, 1989.
  36. E. Nelson. Dynamical theories of Brownian motion, volume 106. Princeton university press, 2020.
  37. The data-driven Schrödinger bridge. Comm. Pure Appl. Math., 74:1545–1573, 2021.
  38. E. Schrödinger. Über die umkehrung der naturgesetze. Sitzungsberichte Preuss. Akad. Wiss. Berlin. Phys. Math., 144:144–153, 1931.
  39. E. Schrödinger. Sur la théorie relativiste de l’électron et l’interprétation de la mécanique quantique. (French) Ann. Inst. H. Poincaré, 2:269–310, 1932.
  40. Score-based generative modeling through stochastic differential equations. arXiv:2011.13456[cs.LG].
  41. Hilbert space embeddings and metrics on probability measures. J. Mach. Learn. Res., 11:1517–1561, 2010.
  42. Solving Schrödinger bridges via maximum likelihood. Entropy, 23:1134, 2021.
  43. H. Wendland. Scattered data approximation. Cambridge University Press, Cambridge, 2010.
  44. J. Yong and X. Y. Zhou. Stochastic controls: Hamiltonian systems and HJB equations. Springer-Verlag, New York, 1999.
  45. J.-C. Zambrini. Variational processes and stochastic versions of mechanics. J. Math. Phys., 27(9):2307–2330, 1986.
Citations (1)
View on Semantic Scholar

Summary

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

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