Location via proxy:   [ UP ]  
[Report a bug]   [Manage cookies]                
Skip to main content

Spherical Spin Glass Model with External Field

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

Abstract

We analyze the free energy and the overlaps in the 2-spin spherical Sherrington Kirkpatrick spin glass model with an external field for the purpose of understanding the transition between this model and the one without an external field. We compute the limiting values and fluctuations of the free energy as well as three types of overlaps in the setting where the strength of the external field goes to zero as the dimension of the spin variable grows. In particular, we consider overlaps with the external field, the ground state, and a replica. Our methods involve a contour integral representation of the partition function along with random matrix techniques. We also provide computations for the matching between different scaling regimes. Finally, we discuss the implications of our results for susceptibility and for the geometry of the Gibbs measure. Some of the findings of this paper are confirmed rigorously by Landon and Sosoe in their recent paper which came out independently and simultaneously.

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

Access this article

Subscribe and save

Springer+ Basic
$34.99 /Month
  • Get 10 units per month
  • Download Article/Chapter or eBook
  • 1 Unit = 1 Article or 1 Chapter
  • Cancel anytime
Subscribe now

Buy Now

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

Instant access to the full article PDF.

Fig. 1
Fig. 2
Fig. 3
Fig. 4
Fig. 5
Fig. 6
Fig. 7
Fig. 8

Similar content being viewed by others

References

  1. Abramowitz, M., Stegun, I.A.: editors. Handbook of mathematical functions with formulas, graphs, and mathematical tables. Dover Publications, Inc., New York, (1992). Reprint of the 1972 edition

  2. Aizenman, M., Lebowitz, J.L., Ruelle, D.: Some rigorous results on the Sherrington–Kirkpatrick spin glass model. Commun. Math. Phys. 112(1), 3–20 (1987)

    Article  ADS  MathSciNet  Google Scholar 

  3. Bai, Z., Yao, J.: On the convergence of the spectral empirical process of Wigner matrices. Bernoulli 11(6), 1059–1092 (2005)

    Article  MathSciNet  Google Scholar 

  4. Baik, J., Lee, J.O.: Fluctuations of the free energy of the spherical Sherrington–Kirkpatrick model. J. Stat. Phys. 165(2), 185–224 (2016)

    Article  ADS  MathSciNet  Google Scholar 

  5. Baik, J., Lee, J.O.: Fluctuations of the free energy of the spherical Sherrington–Kirkpatrick model with ferromagnetic interaction. Ann. Henri Poincaré 18(6), 1867–1917 (2017)

    Article  ADS  MathSciNet  Google Scholar 

  6. Baik, J., Lee, J.O.: Free energy of bipartite spherical Sherrington–Kirkpatrick model. Ann. Inst. Henri Poincaré Probab. Stat. 56(4), 2897–2934 (2020)

    Article  MathSciNet  Google Scholar 

  7. Baik, J., Lee, J.O., Wu, H.: Ferromagnetic to paramagnetic transition in spherical spin glass. J. Stat. Phys. 173(5), 1484–1522 (2018)

    Article  ADS  MathSciNet  Google Scholar 

  8. Binder, K., Young, A.P.: Spin glasses: experimental facts, theoretical concepts, and open questions. Rev. Mod. Phys. 58(4), 801–976 (1986)

    Article  ADS  Google Scholar 

  9. Chen, W.-K., Dey, P., Panchenko, D.: Fluctuations of the free energy in the mixed \(p\)-spin models with external field. Probab. Theory Relat. Fields 168(1–2), 41–53 (2017)

    Article  MathSciNet  Google Scholar 

  10. Chen, W.-K., Sen, A.: Parisi formula, disorder chaos and fluctuation for the ground state energy in the spherical mixed p-spin models. Commun. Math. Phys. 350(1), 129–173 (2016)

    Article  ADS  MathSciNet  Google Scholar 

  11. Collins-Woodfin, E.: Overlap of a spherical spin glass model with microscopic external field. arXiv preprint, arXiv:2101.07704, (2021)

  12. Comets, F., Neveu, J.: The Sherrington–Kirkpatrick model of spin glasses and stochastic calculus: the high temperature case. Commun. Math. Phys. 166(3), 549–564 (1995)

    Article  ADS  MathSciNet  Google Scholar 

  13. Crisanti, A., Sommers, H.-J.: The spherical \(p\)-spin interaction spin glass model: the statics. Zeitschrift für Physik B 87(3), 341–354 (1992)

    Article  ADS  Google Scholar 

  14. Cugliandolo, L.F., Dean, D.S., Yoshino, H.: Nonlinear susceptibilities of spherical models. J. Phys. A 40(16), 4285 (2007)

    Article  ADS  MathSciNet  Google Scholar 

  15. Dembo, A., Zeitouni, O.: Matrix optimization under random external fields. J. Stat. Phys. 159(6), 1306–1326 (2015)

    Article  ADS  MathSciNet  Google Scholar 

  16. Erdős, L., Knowles, A., Yau, H.-T., Yin, J.: Spectral statistics of Erdös-Rényi graphs I: local semicircle law. Ann. Probab. 41(3B), 2279–2375 (2013)

    Article  MathSciNet  Google Scholar 

  17. Erdős, L., Yau, H.-T., Yin, J.: Rigidity of eigenvalues of generalized Wigner matrices. Adv. Math. 229(3), 1435–1515 (2012)

    Article  MathSciNet  Google Scholar 

  18. Fröhlich, J., Zegarliński, B.: Some comments on the Sherrington–Kirkpatrick model of spin glasses. Commun. Math. Phys. 112(4), 553–566 (1987)

    Article  ADS  MathSciNet  Google Scholar 

  19. Fyodorov, Y.V., le Doussal, P.: Topology trivialization and large deviations for the minimum in the simplest random optimization. J. Stat. Phys. 154(1–2), 466–490 (2014)

    Article  ADS  MathSciNet  Google Scholar 

  20. Fyodorov, Y.V., Tublin, R.: Counting stationary points of the loss function in the simplest constrained least-square optimization. Acta Phys. Pol. B 51, 1663–1672 (2020)

    Article  ADS  MathSciNet  Google Scholar 

  21. Guerra, F., Toninelli, F.L.: The thermodynamic limit in mean field spin glass models. Commun. Math. Phys. 230(1), 71–79 (2002)

    Article  ADS  MathSciNet  Google Scholar 

  22. Johansson, K.: On fluctuations of eigenvalues of random Hermitian matrices. Duke Math. J. 91(1), 151–204 (1998)

    Article  MathSciNet  Google Scholar 

  23. Kivimae, P.: Critical fluctuations for the spherical Sherrington-Kirkpatrick model in an external field. arXiv preprint, arXiv:1908.07512 (2019)

  24. Kosterlitz, J., Thouless, D., Jones, R.C.: Spherical model of a spin-glass. Phys. Rev. Lett. 36(20), 1217 (1976)

    Article  ADS  Google Scholar 

  25. Landon, B., Sosoe, P.: Fluctuations of the overlap at low temperature in the 2-spin spherical SK model. arXiv preprint arXiv:1905.03317 (2019)

  26. Landon, B., Sosoe, P.: Fluctuations of the 2-spin SSK model with magnetic field. arXiv preprint arXiv:2009.12514 (2020)

  27. Le Doussal, P., Müller, M., Wiese, K.J.: Cusps and shocks in the renormalized potential of glassy random manifolds: how functional renormalization group and replica symmetry breaking fit together. Phys. Rev. B 77, 064203 (2008)

    Article  ADS  Google Scholar 

  28. Le Doussal, P., Müller, M., Wiese, K.J.: Avalanches in mean-field models and the Barkhausen noise in spin-glasses. Europhys. Lett. 91(5), 57004 (2010)

    Article  ADS  Google Scholar 

  29. Le Doussal, P., Müller, M., Wiese, K.J.: Equilibrium avalanches in spin glasses. Phys. Rev. B 85, 214402 (2012)

    Article  ADS  Google Scholar 

  30. Lytova, A., Pastur, L.: Central limit theorem for linear eigenvalue statistics of random matrices with independent entries. Ann. Probab. 37(5), 1778–1840 (2009)

    Article  MathSciNet  Google Scholar 

  31. Mehta, M.L.: Random matrices, volume 142 of Pure and Applied Mathematics (Amsterdam). Elsevier/Academic Press, Amsterdam, third edition, (2004)

  32. Mézard, M., Parisi, G., Virasoro, M.A.: Spin Glass Theory and Beyond. World Scientific Lecture Notes in Physics. World Scientific Publishing Co., Inc, Teaneck (1987)

    MATH  Google Scholar 

  33. Nguyen, V.L., Sosoe, P.: Central limit theorem near the critical temperature for the overlap in the 2-spin spherical SK model. J. Math. Phys. 60(10), 103302–103313 (2019)

    Article  ADS  MathSciNet  Google Scholar 

  34. Panchenko, D., Talagrand, M.: On the overlap in the multiple spherical SK models. Ann. Probab. 35(6), 2321–2355 (2007)

    Article  MathSciNet  Google Scholar 

  35. Parisi, G.: The order parameter for spin glasses: a function on the interval 0–1. J. Phys. A 13(3), 1101–1112 (1980)

    Article  ADS  Google Scholar 

  36. Parisi, G.: A sequence of approximated solutions to the SK model for spin glasses. J. Phys. A 13(4), L115 (1980)

    Article  ADS  Google Scholar 

  37. Sherrington, D., Kirkpatrick, S.: Solvable model of a spin-glass. Phys. Rev. Lett. 35(26), 1792–1796 (1975)

    Article  ADS  Google Scholar 

  38. Soshnikov, A.: Universality at the edge of the spectrum in Wigner random matrices. Commun. Math. Phys. 207(3), 697–733 (1999)

    Article  ADS  MathSciNet  Google Scholar 

  39. Subag, E.: The geometry of the Gibbs measure of pure spherical spin glasses. Invent. Math. 210(1), 135–209 (2017)

    Article  ADS  MathSciNet  Google Scholar 

  40. Talagrand, M.: Replica symmetry breaking and exponential inequalities for the Sherrington–Kirkpatrick model. Ann. Probab. 28(3), 1018–1062 (2000)

    Article  MathSciNet  Google Scholar 

  41. Talagrand, M.: Free energy of the spherical mean field model. Probab. Theory Relat. Fields 134(3), 339–382 (2006)

    Article  MathSciNet  Google Scholar 

  42. Talagrand, M.: The Parisi formula. Ann. Math. 163(1), 221–263 (2006)

    Article  MathSciNet  Google Scholar 

  43. Tracy, C.A., Widom, H.: Level-spacing distributions and the Airy kernel. Commun. Math. Phys. 159(1), 151–174 (1994)

    Article  ADS  MathSciNet  Google Scholar 

  44. Yoshino, H., Rizzo, T.: Stepwise responses in mesoscopic glassy systems: a mean-field approach. Phys. Rev. B 77, 104429 (2008)

    Article  ADS  Google Scholar 

  45. Young, A.P.: Spin Glasses and Random Fields, vol. 12. World Scientific, Singapore (1998)

    Google Scholar 

  46. Young, A.P., Bray, A.J., Moore, M.A.: Lack of self-averaging in spin glasses. J. Phys. 17(5), L149–L154 (1984)

    ADS  Google Scholar 

Download references

Acknowledgements

The authors would like to thank Benjamin Landon and Philippe Sosoe for sharing their recent work with us. The work of Baik was supported in part by the NSF grants DMS-1664692 and DMS-1954790. The work of Collins-Woodfin was supported in part by the NSF grants DMS-1701577 and DMS-1954790. The work of Le Doussal was supported in part by the ANR grant ANR-17- CE30-0027-01 RaMaTraF. Le Doussal would like to thank the Department of Mathematics of the University of Michigan for hospitality; this joint project started during his visit to the department.

Author information

Authors and Affiliations

Authors

Corresponding author

Correspondence to Jinho Baik.

Additional information

Communicated by Federico Ricci-Tersenghi.

Publisher's Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Appendices

Appendix A Proof of Lemma 3.3

We prove Lemma 3.3. First,

$$\begin{aligned} \langle e^{\beta \eta \mathfrak {M}}\rangle = \frac{1}{\mathcal {Z}_N(h)} \int _{S_{N - 1}} e^{\beta \frac{ \eta }{N} \mathbf{g }\cdot \varvec{\sigma }} e^{\beta \left( \frac{1}{2} \varvec{\sigma }\cdot M\varvec{\sigma }+h \mathbf{g }\cdot \varvec{\sigma }\right) } \mathrm {d}\omega _N(\varvec{\sigma }) = \frac{\mathcal {Z}_N(h+ \eta N^{-1})}{\mathcal {Z}_N(h)}. \end{aligned}$$

Secondly, by definition,

$$\begin{aligned} \langle e^{\beta \eta \mathfrak {O}} \rangle = \frac{1}{\mathcal {Z}_N} \int _{S_{N - 1}} e^{\beta \frac{\eta }{N} (\mathbf{u }_1 \cdot \sigma )^2} e^{\beta \left( \varvec{\sigma }\cdot M\varvec{\sigma }+h \mathbf{g }\cdot \varvec{\sigma }\right) } \mathrm {d}\omega _N(\varvec{\sigma }). \end{aligned}$$
(A.1)

Since

$$\begin{aligned} \frac{1}{2} \varvec{\sigma }\cdot M\varvec{\sigma }+ \frac{\eta }{N} (\mathbf{u }_1\cdot \sigma )^2= \frac{1}{2} \sum _{i=1}^N \lambda _i (\mathbf{u }_i\cdot \sigma )^2+ \frac{\eta }{N} (\mathbf{u }_1\cdot \sigma )^2, \end{aligned}$$

the integral in (A.1) is the same as that of \(\mathcal {Z}_N\) with \(\lambda _1\mapsto \lambda _1+ \frac{2\eta }{N}\). Finally, using the eigenvalue-eigenvector decomposition \(M=O\Lambda O^T\) and changing variables \(\frac{1}{\sqrt{N}} O^T\sigma =x\) and \(\frac{1}{\sqrt{N}} O^T \tau =y\), we find that

$$\begin{aligned} \langle e^{\eta \mathfrak {R}} \rangle = \frac{J(\frac{\beta N}{2}, \frac{\beta N}{2}; \frac{\eta }{N \beta }, \frac{\sqrt{\beta }h}{\sqrt{2}}, \frac{\sqrt{\beta }h}{\sqrt{2}})}{J(\frac{\beta N}{2}, \frac{\beta N}{2}; 0, \frac{\sqrt{\beta }h}{\sqrt{2}}, \frac{\sqrt{\beta }h}{\sqrt{2}})}. \end{aligned}$$
(A.2)

where we use the notation

$$\begin{aligned}&J(u,v; a,b,c) \\&\quad = (uv)^{\frac{N}{2} - 1}\displaystyle {\int \int } e^{ 2a\sqrt{uv}\sum \limits _{i = 1}^N x_i y_i +u \sum \limits _{i=1}^N \lambda _i x_i^2 + 2b\sqrt{u} \sum \limits _{i=1}^N n_i x_i + v \sum \limits _{i=1}^N \lambda _i y_i^2 + 2c\sqrt{v} \sum \limits _{i=1}^N n_i y_i } \mathrm {d}\Omega ^{\otimes 2}_{N-1}(x, y) . \end{aligned}$$

We evaluate the Laplace transform of J(uvabc). Changing of variable as \(u = r^2\), \(v = s^2\) and \(rx \mapsto x\), \(sy \mapsto y\), the Laplace transform

$$\begin{aligned} Q(z,w) = \int _0^\infty \int _0^\infty e^{-zu - wv} J(u,v) \mathrm {d}u \mathrm {d}v \end{aligned}$$

becomes a 2-dimensional Gaussian integral which evaluates to

$$\begin{aligned} Q(z,w) = 4 \prod \limits _{i = 1}^N \frac{\pi }{\sqrt{(z - \lambda _i)(w - \lambda _i) - a^2}} e^{\frac{n_i^2((w - \lambda _i)b^2 + 2abc + (z - \lambda _i) c^2}{(z - \lambda _i)(w - \lambda _i) - a^2}}. \end{aligned}$$

The inverse Laplace transform gives a double integral formula for J(uv).

Appendix B A perturbation argument

The following perturbation lemma is used to obtain (5.7), (5.22) and (6.4).

Lemma B.1

Let I be a closed interval of \({\mathbb {R}}\). Let G(zN) be a sequence of random \(C^{4}\)-functions for \(z \in I\). Let \(\epsilon = \epsilon (N) := N^{-\delta }\) for some \(\delta > 0\) and assume that

$$\begin{aligned} G(z; N)= G_0(z; N) + G_1(z; N) \epsilon + G_2(z; N) \epsilon ^2 + {\mathcal {O}}\left( \epsilon ^3\right) \end{aligned}$$
(B.1)

and

$$\begin{aligned} G'(z; N) = G'_0(z; N) + G'_1(z; N) \epsilon + G'_2(z; N) \epsilon ^2 + {\mathcal {O}}\left( \epsilon ^3\right) \end{aligned}$$
(B.2)

for random \(C^{4}\)-functions \(G_k(z; N)\). Suppose that

$$\begin{aligned} G_k^{(\ell )}(z;N) = {\mathcal {O}}\left( 1\right) \end{aligned}$$
(B.3)

uniformly for \(z\in I\) for all \(k=0,1,2\), \(0\le \ell \le 4\) and also assume that there is a \(\gamma _0 \in I\) satisfying

$$\begin{aligned} G_0'(\gamma _0;N)=0, \qquad |G_0''(\gamma _0;N) |\ge C > 0 \end{aligned}$$
(B.4)

for a positive constant C. Then there is a critical point \(\gamma =\gamma (N)\) of G(zN) admitting the asymptotic expansion

$$\begin{aligned} \gamma = \gamma _0+ \gamma _1 \epsilon + \gamma _2 \epsilon ^2 + {\mathcal {O}}\left( \epsilon ^3\right) \end{aligned}$$
(B.5)

where

$$\begin{aligned} \gamma _1 = -\frac{G_1'(\gamma _0;N)}{G_0''(\gamma _0;N)}, \qquad \gamma _2 = -\frac{G_2'(\gamma _0;N) + G_1''(\gamma _0;N)\gamma _1 + \frac{1}{2} G_0'''(\gamma _0;N)\gamma _1^2}{G_0''(\gamma _0;N)}. \end{aligned}$$
(B.6)

Furthermore,

$$\begin{aligned} \begin{aligned} G(\gamma ;N) = G_0(\gamma _0;N) + G_1(\gamma _0;N) \epsilon + \left( \frac{1}{2} G_1'(\gamma _0;N)\gamma _1 + G_2(\gamma _0;N) \right) \epsilon ^2 + {\mathcal {O}}\left( \epsilon ^3\right) . \end{aligned} \end{aligned}$$
(B.7)

Proof

This lemma is standard when G(zN) is deterministic. The proof for the random G(zN) does not change. For simplicity, we suppress the dependence on N in the notations; for example we write \(G_0(z)\) instead of \(G_0(z;N)\). In order to prove (B.5), it is enough to show that for any \(0< t < \delta \), \(G'(\gamma _+)G'(\gamma _-) < 0\) with \(\gamma _\pm = \gamma _0 + \gamma _1 \epsilon + \gamma _2 \epsilon ^2 \pm \epsilon ^3N^t\). From the Taylor expansion,

$$\begin{aligned} \begin{aligned} G'(\gamma _\pm )&= G'_0(\gamma _0) + (G_0''(\gamma )\gamma _1 + G_1'(\gamma _0)) \epsilon \\&\quad + \left( G_0''(\gamma _0)\gamma _2 + G_2'(\gamma _0) + G_1''(\gamma _0)\gamma _1 + \frac{1}{2} G_0'''(\gamma _0)\gamma _1^2 \right) \epsilon ^2 \\&\quad \pm G''_0(\gamma _0) \epsilon ^3 N^t + {\mathcal {O}}\left( \epsilon ^3\right) . \end{aligned} \end{aligned}$$
(B.8)

The definitions of \(\gamma _0, \gamma _1\), and \(\gamma _2\) imply that

$$\begin{aligned} \begin{aligned} G'(\gamma _\pm )&= \pm G''_0(\gamma _0) \epsilon ^3 N^t + {\mathcal {O}}\left( \epsilon ^3\right) \end{aligned} \end{aligned}$$
(B.9)

Thus, \(G'(\gamma _+)G'(\gamma _-) <0\) for all large enough N and we obtain (B.5). The Eq. (B.7) follows from

$$\begin{aligned} \begin{aligned} G(\gamma )&= G_0(\gamma ) + G_1(\gamma ) \epsilon + G_2(\gamma ) \epsilon ^2 + {\mathcal {O}}\left( \epsilon ^3\right) = G_0(\gamma _0) + (G_0'(\gamma _0)\gamma _1 + G_1(\gamma _0)) \epsilon \\&\quad + \left( G_0'(\gamma _0)\gamma _2 + \frac{1}{2} G_0''(\gamma _0)\gamma _1^2 + G_1'(\gamma _0)\gamma _1 + G_2(\gamma _0) \right) \epsilon ^2 + {\mathcal {O}}\left( \epsilon ^3\right) , \end{aligned} \end{aligned}$$
(B.10)

together with \(G'_0(\gamma _0) = 0\) and (B.6). \(\square \)

Remark B.2

Here, we consider the asymptotic expansion of G(z) up to the third order term. One can also consider the case where the expansion is up to the second order, then (B.7) is still valid up to the second order.

Rights and permissions

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Baik, J., Collins-Woodfin, E., Le Doussal, P. et al. Spherical Spin Glass Model with External Field. J Stat Phys 183, 31 (2021). https://doi.org/10.1007/s10955-021-02757-7

Download citation

  • Received:

  • Accepted:

  • Published:

  • DOI: https://doi.org/10.1007/s10955-021-02757-7

Keywords