Skip to main content
Log in

The Full Replica Symmetry Breaking in the Ising Spin Glass on Random Regular Graph

  • Published:
Journal of Statistical Physics Aims and scope Submit manuscript

Abstract

In this paper, we extend the full replica symmetry breaking scheme to the Ising spin glass on a random regular graph. We propose a new martingale approach, that overcomes the limits of the Parisi–Mézard cavity method, providing a well-defined formulation of the full replica symmetry breaking problem in random regular graphs. Finally, we define the order parameters of the system and get a set of self-consistency equations for the order parameters and the free energy. We face up the problem only from a technical point of view: the physical meaning of this approach and the quantitative evaluation of the solution of the self-consistency equations will be discussed in next works.

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

Access this article

Price excludes VAT (USA)
Tax calculation will be finalised during checkout.

Instant access to the full article PDF.

Similar content being viewed by others

Notes

  1. In the original works by Austin and Panchenko [38, 39], a random array is generated by a function of uniform random variables on [0, 1]. A uniform random variable, however, can be generated in distribution as a function of a Gaussian variable, than the representation presented here is equivalent to the Austin and Panchenko representation.

  2. The usual completion of a continuous-time filtration, with respect to a given probability measure, is the smallest right-continuous filtration that contains the original one, enlarged with the set with probability 0 with respect to the given probability measure. Usually, a rigorous treatment of continuous-time stochastic processes requires the usual completion.

References

  1. Sherrington, D., Kirkpatrick, S.: Solvable model of a spin-glass. Phys. Rev. Lett. 35, 1792 (1975)

    Article  ADS  Google Scholar 

  2. Kirkpatrick, S., Sherrington, D.: Infinite-ranged models of spin-glasses. Phys. Rev. B 17, 4384 (1978)

    Article  ADS  Google Scholar 

  3. Edwards, S., Anderson, P.: Theory of spin glasses. J. Phys. F 6, 1927 (1976)

    Article  ADS  Google Scholar 

  4. Parisi, G.: Toward a mean-field theory for spin glasses. Phys. Lett. 73A, 203 (1979)

    Article  ADS  MathSciNet  Google Scholar 

  5. Parisi, G.: A sequence of approximated solutions to the S-K model for spin glasses. J. Phys. A 13, L115 (1980)

    Article  ADS  Google Scholar 

  6. Parisi, G.: The order parameter for spin glasses: a function on the interval 0–1. J. Phys. A 13, 1101 (1980)

    Article  ADS  Google Scholar 

  7. Mézard, M., Parisi, G., Virasoro, M.A.: Spin Glass Theory and Beyond. World Scientific, Singapore (1987)

    MATH  Google Scholar 

  8. Parisi, G.: Order parameter for spin-glasses. Phys. Rev. Lett. A 50, 1946 (1983)

    Article  ADS  MathSciNet  Google Scholar 

  9. Nishimori, H.: Statistical Physics of Spin Glasses and Information Processing: An Introduction. Oxford University Press, Oxford (2001)

    Book  Google Scholar 

  10. Derrida, B.: Random-energy model: an exactly solvable model of disordered systems. Phys. Rev. B 24, 2613 (1981)

    Article  ADS  MathSciNet  Google Scholar 

  11. Crisanti, A., Sommers, H.-J.: The spherical p-spin interaction spin glass model: the statics. J. Phys. B 87, 341 (1992)

    ADS  Google Scholar 

  12. Virasoro, A., Mézard, M.: The microstructure of ultrametricity. J. Phys. 46, 1293 (1985)

    Article  MathSciNet  Google Scholar 

  13. Derrida, B.: A generalization of the random energy model which includes correlations between energies. J. Phys. Lett. 46, L401 (1985)

    Article  ADS  Google Scholar 

  14. Mézard, M., Parisi, G., Virasoro, A.: Random free energies in spin glasses. J. Phys. Lett. 46, L217 (1985)

    Article  Google Scholar 

  15. Ruelle, D.: A mathematical reformulation of Derridas REM and GREM. Commun. Math. Phys. 108, 225 (1987)

    Article  ADS  MathSciNet  Google Scholar 

  16. Aizenman, M., Sims, R., Starr, S.L.: Extended variational principle for the Sherrington–Kirkpatrick spin-glass model. Phys. Rev. B 68, 214403 (2003)

    Article  ADS  Google Scholar 

  17. Arguin, L.-P.: Spin glass computations and Ruelles probability cascades. J. Stat. Phys. 126, 951 (2007)

    Article  ADS  MathSciNet  Google Scholar 

  18. De Dominicis, C., Kondor, I.: Eigenvalues of the stability matrix for Parisi solution of the long-range spin-glass. Phys. Rev. B 27, 606 (1983)

    Article  ADS  MathSciNet  Google Scholar 

  19. Talagrand, M.: The Parisi formula. Ann. Math. 163, 221 (2006)

    Article  MathSciNet  Google Scholar 

  20. Guerra, F.: Broken replica symmetry bounds in the mean-field spin glass model. Commun. Math. Phys. 233, 1 (2003)

    Article  ADS  MathSciNet  Google Scholar 

  21. Panchenko, D.: Ghirlanda Guerra identities and ultrametricity: an elementary proof in the discrete case. C. R. Acad. Sci. Paris Ser. I 349, 813 (2011)

    Article  MathSciNet  Google Scholar 

  22. Panchenko, D.: The Parisi ultrametricity conjecture. Ann. Math. 177, 383 (2013)

    Article  MathSciNet  Google Scholar 

  23. Ghirlanda, S., Guerra, F.: General properties of overlap probability distributions in disordered spin systems. Towards Parisi ultrametricity. J. Phys. A 31, 9149 (1998)

    Article  ADS  MathSciNet  Google Scholar 

  24. Bollobás, B.: Random Graphs. Cambridge University Press, Cambridge (2001)

    Book  Google Scholar 

  25. Mézard, M., Parisi, G.: The Bethe lattice spin glass revisited. Eur. Phys. J. B 20, 217 (2001)

    Article  ADS  MathSciNet  Google Scholar 

  26. Mézard, M., Parisi, G., Zecchina, R.: Analytic and algorithmic solution of random satisfiability problems. Science 297, 812 (2002)

    Article  ADS  Google Scholar 

  27. Mézard, M., Parisi, G.: The cavity method at zero temperature. J. Stat. Phys. 111, 1 (2003)

    Article  MathSciNet  Google Scholar 

  28. Mézard, M., Montanari, A.: Information, Physics, and Computation. Oxford University Press, Oxford (2009)

    Book  Google Scholar 

  29. Panchenko, D.: Spin glass models from the point of view of spin distributions. Ann. Probab. 41, 1315 (2013)

    Article  MathSciNet  Google Scholar 

  30. Panchenko, D.: Structure of finite-RSB asymptotic Gibbs measures in the diluted spin glass models. J. Stat. Phys. 162, 1 (2016)

    Article  ADS  MathSciNet  Google Scholar 

  31. Franz, S., Leone, M.: Replica bounds for optimization problems and diluted spin systems. J. Stat. Phys. 111, 535 (2003)

    Article  MathSciNet  Google Scholar 

  32. Franz, S., Leone, M., Toninelli, F.L.: Replica bounds for diluted non-Poissonian spin systems. J. Phys. A 43, 10967 (2003)

    Article  MathSciNet  Google Scholar 

  33. Parisi, G.: The marginally stable bethe lattice spin glass revisite. J. Stat. Phys. 167, 515 (2017)

    Article  ADS  MathSciNet  Google Scholar 

  34. Revuz, D., Yor, D.: Continuous Martingales and Brownian Motion. Springer, Berlin (1999)

    Book  Google Scholar 

  35. Boué, M., Dupuis, P.: A variational representation for certain functiona of Brownian motion. Ann. Probab. 26, 1641 (1998)

    Article  MathSciNet  Google Scholar 

  36. Auffinger, A., Chen, W.-K.: The Parisi formula has a unique minimizer. Commun. Math. Phys. 335, 1429 (2015)

    Article  ADS  MathSciNet  Google Scholar 

  37. Panchenko, D.: Hierarchical exchangeability of pure states in mean field spin glass models. Probab. Theory Relat. Fields 161, 619 (2015)

    Article  MathSciNet  Google Scholar 

  38. Austin, T.: On exchangeable random variables and the statistics of large graphs and hypergraphs. Probab. Surv. 5, 80 (2008)

    Article  MathSciNet  Google Scholar 

  39. Austin, T., Panchenko, D.: A hierarchical version of the de Finetti and Aldous-Hoover representations. Probab. Theory Relat. Fields 159, 809 (2014)

    Article  MathSciNet  Google Scholar 

  40. Billingsley, P.: Probability and Measure. Wiley, New York (1995)

    MATH  Google Scholar 

  41. Dupire, B.: Functional It Calculus. Portfolio Research Paper 2009-04, Bloomberg ( 2009)

  42. Cont, R., Fournizé, A.: Functional Itô calculus and stochastic integral representation of martingales. Ann. Probab. 41(1), 109 (2013)

    Article  MathSciNet  Google Scholar 

  43. Crisanti, A., Rizzo, T.: Analysis of the \(\infty \)-replica symmetry breaking solution of the Sherrington-Kirkpatrick model. Phys. Rev. E 65(4), 046137 (2002)

    Article  ADS  MathSciNet  Google Scholar 

  44. Crisanti, A., Leuzzi, L.: Spherical 2+p spin-glass model: an analytically solvable model with a glass-to-glass transition. Phys. Rev. B 73, 014412 (2006)

    Article  ADS  Google Scholar 

  45. Itô, K.: Stochastic integral. In: Proceedings of Imperial Academy, Japan, Vol. 20, p. 519 (1944)

    Article  MathSciNet  Google Scholar 

  46. Doob, J.L.: Stochastic Processes. Wiley, New York (1953)

    MATH  Google Scholar 

  47. Clark, J.M.C.: The representation of functionals of brownian motion by stochastic integrals. Ann. Math. Stat. 41, 1282 (1970)

    Article  MathSciNet  Google Scholar 

  48. Cameron, R.H., Martin, W.T.: Transformation of Wiener integrals under translations. Ann. Math. 45, 386 (1944)

    Article  MathSciNet  Google Scholar 

  49. Girsanov, I.V.: On transforming a certain class of stochastic processes by absolutely continuous substitution of measures. Theory Probab. Appl. 5, 285 (1960)

    Article  MathSciNet  Google Scholar 

  50. Malliavin, P., Thalmaier, A.: Stochastic Calculus of Variations in Mathematical Finance. Springer, Berlin (2005)

    MATH  Google Scholar 

Download references

Acknowledgements

I am grateful to Giorgio Parisi and Simone Franchini for interesting discussions.

Author information

Authors and Affiliations

Authors

Corresponding author

Correspondence to Francesco Concetti.

Appendix: The 1 RSB Free Energy Functional

Appendix: The 1 RSB Free Energy Functional

In this appendix, we show that the replica symmetry breaking ansatz described in the Sect. 3.1 reproduces the

Let us define the functions

$$\begin{aligned} F^{\text {(v)}}(J_{0,1},\ldots ,J_{0,c},h_1, \ldots ,h_c\,)=\log \Delta ^{\text {(v)}}(J_{0,1}, \ldots ,J_{0,c},h_1,\ldots ,h_c), \end{aligned}$$
(77)

and

$$\begin{aligned} F^{\text {(e)}}(J_{1,2},h_1,h_2\,)= \Delta ^{\text {(e)}}(J_{1,2},h_{1},h_{2}). \end{aligned}$$
(78)

Putting \(r=1\), the cavity field functional (16) is a measurable function of two independent normal random variables \(W^{(0)}\) and \(W^{(1)}\):

$$\begin{aligned} \{W^{(0)},W^{(1)}\}\mapsto h\big (\,W^{(0)},W^{(1)}\,\big ). \end{aligned}$$
(79)

The edge and vertex contributions to the free energy are given by a single iteration of the iterative rule (21):

$$\begin{aligned}&\phi ^{\text {(e/v)}}_{0}\big (\varvec{J},\,\varvec{W}^{(0)}\,\big )=\frac{1}{x_{1}}\log \mathbb {E}_{\mathbb {W}_1^{\otimes (2/c)}}\left[ \,\exp \left( \,x_{1} \phi ^{\text {(e/v)}}_{1}\big (\varvec{J},\varvec{W}^{(0)},\, \varvec{W}^{(1)}\,\big ) \,\right) \,\Big |\varvec{W}^{(0)}\right] \nonumber \\&\quad =\frac{1}{x_{1}}\log \left( \int \left( \prod ^{(2/c)}_{i=1} d\nu \big (W^{(1)}_{i}\big )\,\right) e^{\,x_{1}\, F^{\text {(e/v)}}\left( \varvec{J},\varvec{h}\big (\,\varvec{W}^{(0)},\,\varvec{W}^{(1)}\,\big )\,\,\right) \,}\,\,\right) , \end{aligned}$$
(80)

where \(\nu \) is the normal distribution.

The free energy functional is finally given by

$$\begin{aligned} \Phi =\overline{\int \prod ^c_{i=1} d\nu \big (W_i^{(0)}\,\big )\,\,\frac{1}{x_{1}} \log \left( \int \left( \prod ^{c}_{i=1} d\nu \big (W^{(1)}_{i}\big )\,\right) e^{\,x_{1}\, F^{\text {(v)}}\left( J_{0,1},\ldots ,J_{0,c},\,h_ 1\big (\,W_1^{(0)},\,W_1^{(1)}\,\big ), \ldots \,h_c\big (\,W_c^{(0)}, \,W_c^{(1)}\,\big )\,\right) \,}\,\,\right) }\nonumber \\ -\frac{c}{2}\overline{\int d\nu \big (W_1^{(0)}\big )d\nu \big (W_2^{(0)}\big ) \frac{1}{x_{1}}\log \left( \int _{\mathbb {R}^{(2/c)}}d \nu \big (W_1^{(1)}\big )d\nu \big (W_2^{(1)}\big )\,e^{\,x_{1}\, F^{\text {(e)}}\left( J_{1,2},h_1\big (\,W_1^{(0)}, \,W_1^{(1)}\,\big ),\,h_2\big (\,W_2^{(0)}, \,W_2^{(1)}\,\big )\,\right) \,}\,\,\right) }.\nonumber \\ \end{aligned}$$
(81)

Now, let us define the probability density distribution of the cavity field h, conditionally to a fixed value for the random variable \(W^{(0)}\):

$$\begin{aligned}&\widehat{\pi }\big (y\big |W^{(0)}\,\big )=\mathbb {E}_{\mathbb {W}_1^{\otimes (2/c)}}\left[ \delta \big (\,y-h\big (\,W^{(0)},W^{(1)}\,\big )\,\big )\Big |W^{(0)}\right] \nonumber \\&\quad =\int d\nu \big (W^{(1)}\big )\delta \big (\,h\big (\,W^{(0)},W^{(1)}\,\big )-y\,\big ),\quad y\in \mathbb {R}, \end{aligned}$$
(82)

where the symbol \(\delta (\cdot )\) denotes the Dirac delta distribution. For each fixed value of \(y\in \mathbb {R}\), the quantity \(\widehat{\pi }(y|\,\cdot \,)\) is a positive random variable, since it depends on \(W^{(0)}\). Then, we can define the probability density distribution of the random probability density distribtion \(\widehat{\pi }:=\big \{\widehat{\pi }\big (y\big |\,\cdot \,\big );\,y\in \mathbb {R}\,\big \}\) in such a way:

$$\begin{aligned} \Pi [\,\pi \,]=\mathbb {E}_{\mathbb {W}_0^{\otimes (2/c)}}\left[ \delta \big [\,\pi \big (\,\cdot \,\big )-\widehat{\pi }\big (\,\cdot \,\big |W^{(0)}\,\big )\,\big ]\,\right] =\int d\nu \big (W^{(0)}\big )\delta \big [\,\pi \big (\,\cdot \,\big )-\widehat{\pi }\big (\,\cdot \,\big |W^{(0)}\,\big )\,\big ], \end{aligned}$$
(83)

where \(\delta [\,\cdot \,]\) is the functional Dirac delta. Substituting the Eqs. (82) and (83) in(81), one get

$$\begin{aligned}&x_1 \Phi =x_1 \Phi \big [\Pi ,\pi ,x_1\big ]=\overline{\int \left( \prod ^c_{i=1} d[\pi _i]\Pi [\,\pi _i\,]\right) \,\,\log \left( \int \left( \prod ^{c}_{i=1} dy \,\pi \big (y\big )\,\right) e^{\,x_{1}\, F^{\text {(v)}}\left( J_{0,1},\ldots ,J_{0,c}, \,y_1,\ldots \,y_c\,\big )\,\right) }\right) }\nonumber \\&\quad -\frac{c}{2}\overline{\int d[\pi _1]\Pi [\,\pi _1\,]\,\,d[\pi _2]\Pi [\,\pi _2\,] \log \left( \int _{\mathbb {R}^{(2/c)}}\,dy_1\,\pi _i(y_1)\,dy_2\,\pi _i(y_2)\,e^{\,x_{1}\, F^{\text {(e)}}\left( J_{1,2},y_1,\,y_2\right) }\right) }, \end{aligned}$$
(84)

where the symbol \(\int d[\pi ]\) represents the functional integral.

The variational free energy functional (84) is equivalent to a finite temperature version of the 1-RSB variational free energy presented in Appendix B of [27]. The functional \(\Pi \) is the 1-RSB cavity method order parameter and \(x_1\) is the Parisi 1-RSB parameter.

Rights and permissions

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Concetti, F. The Full Replica Symmetry Breaking in the Ising Spin Glass on Random Regular Graph. J Stat Phys 173, 1459–1483 (2018). https://doi.org/10.1007/s10955-018-2142-6

Download citation

  • Received:

  • Accepted:

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1007/s10955-018-2142-6

Keywords

Navigation