Machine Learning of Explicit Order Parameters: From the Ising Model to SU(2) Lattice Gauge Theory

We present a solution to the problem of interpreting neural networks classifying phases of matter. We devise a procedure for reconstructing the decision function of an artificial neural network as a simple function of the input, provided the decision function is sufficiently symmetric. In this case one can easily deduce the quantity by which the neural network classifies the input. The method is applied to the Ising model and SU(2) lattice gauge theory. In both systems we deduce the explicit expressions of the order parameters from the decision functions of the neural networks. We assume no prior knowledge about the Hamiltonian or the order parameters except Monte Carlo–sampled configurations.

[1]  Anna Levit,et al.  Reinforcement learning using quantum Boltzmann machines , 2016, Quantum Inf. Comput..

[2]  Roger G. Melko,et al.  Kernel methods for interpretable machine learning of order parameters , 2017, 1704.05848.

[3]  Titus Neupert,et al.  Probing many-body localization with neural networks , 2017, 1704.01578.

[4]  Wenjian Hu,et al.  Discovering phases, phase transitions, and crossovers through unsupervised machine learning: A critical examination. , 2017, Physical review. E.

[5]  Sebastian Johann Wetzel,et al.  Unsupervised learning of phase transitions: from principal component analysis to variational autoencoders , 2017, Physical review. E.

[6]  Naftali Tishby,et al.  Opening the Black Box of Deep Neural Networks via Information , 2017, ArXiv.

[7]  G. Kasieczka,et al.  Deep-learning top taggers or the end of QCD? , 2017, 1701.08784.

[8]  Lu-Ming Duan,et al.  Efficient representation of quantum many-body states with deep neural networks , 2017, Nature Communications.

[9]  D. Deng,et al.  Quantum Entanglement in Neural Network States , 2017, 1701.04844.

[10]  Isaac Tamblyn,et al.  Sampling algorithms for validation of supervised learning models for Ising-like systems , 2017, J. Comput. Phys..

[11]  Yi Zhang,et al.  Quantum Loop Topography for Machine Learning. , 2016, Physical review letters.

[12]  Li Huang,et al.  Accelerated Monte Carlo simulations with restricted Boltzmann machines , 2016, 1610.02746.

[13]  S. Huber,et al.  Learning phase transitions by confusion , 2016, Nature Physics.

[14]  R. Melko,et al.  Machine Learning Phases of Strongly Correlated Fermions , 2016, Physical Review X.

[15]  Juan Carrasquilla,et al.  Machine learning quantum phases of matter beyond the fermion sign problem , 2016, Scientific Reports.

[16]  K. Aoki,et al.  Restricted Boltzmann machines for the long range Ising models , 2016, 1701.00246.

[17]  Dong-Ling Deng,et al.  Exact Machine Learning Topological States , 2016 .

[18]  Kieron Burke,et al.  Pure density functional for strong correlation and the thermodynamic limit from machine learning , 2016, 1609.03705.

[19]  Roger G. Melko,et al.  Learning Thermodynamics with Boltzmann Machines , 2016, ArXiv.

[20]  Lei Wang,et al.  Discovering phase transitions with unsupervised learning , 2016, 1606.00318.

[21]  Roger G. Melko,et al.  Machine learning phases of matter , 2016, Nature Physics.

[22]  Demis Hassabis,et al.  Mastering the game of Go with deep neural networks and tree search , 2016, Nature.

[23]  Zoubin Ghahramani,et al.  Probabilistic machine learning and artificial intelligence , 2015, Nature.

[24]  Andrea Vedaldi,et al.  Understanding deep image representations by inverting them , 2014, 2015 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).

[25]  Geoffrey E. Hinton,et al.  ImageNet classification with deep convolutional neural networks , 2012, Commun. ACM.

[26]  Tara N. Sainath,et al.  Deep Neural Networks for Acoustic Modeling in Speech Recognition: The Shared Views of Four Research Groups , 2012, IEEE Signal Processing Magazine.

[27]  Gaël Varoquaux,et al.  Scikit-learn: Machine Learning in Python , 2011, J. Mach. Learn. Res..

[28]  Simon Friederich,et al.  Functional renormalization for spontaneous symmetry breaking in the Hubbard model , 2010, 1012.5436.

[29]  S. Raghu,et al.  Superconductivity from repulsive interactions in the two-dimensional electron gas , 2010, 1009.3600.

[30]  L. Taillefer.  Scattering and Pairing in Cuprate Superconductors , 2010, 1003.2972.

[31]  A. Damascelli,et al.  Two gaps make a high-temperature superconductor? , 2007, 0706.4282.

[32]  C. Varma.  Theory of the pseudogap state of the cuprates , 2005, cond-mat/0507214.

[33]  P. Atzberger.  The Monte-Carlo Method , 2006 .

[34]  D. Rischke.  The Phases of Quantum Chromodynamics: From Confinement to Extreme Environments , 2005 .

[35]  P. Kent,et al.  Systematic study of d-wave superconductivity in the 2D repulsive Hubbard model. , 2005, Physical review letters.

[36]  M. Stephanov.  QCD Phase Diagram and the Critical Point , 2005, hep-ph/0402115.

[37]  H. Ren.  Color Superconductivity in a Dense Quark Matter , 2003, hep-ph/0307125.

[38]  S. Uchida,et al.  High field phase diagram of cuprates derived from the Nernst effect. , 2002, Physical review letters.

[39]  Jude W. Shavlik,et al.  Interpretation of Artificial Neural Networks: Mapping Knowledge-Based Neural Networks into Rules , 1991, NIPS.

[40]  Kurt Hornik,et al.  Approximation capabilities of multilayer feedforward networks , 1991, Neural Networks.

[41]  George Cybenko,et al.  Approximation by superpositions of a sigmoidal function , 1989, Math. Control. Signals Syst..

[42]  Michael Creutz,et al.  Monte Carlo Study of Quantized SU(2) Gauge Theory , 1980 .

[43]  Kenneth G. Wilson,et al.  Quantum Chromodynamics on a Lattice , 1977 .

[44]  L. Onsager.  Crystal statistics. I. A two-dimensional model with an order-disorder transition , 1944 .

[45]  Karl Pearson F.R.S..  LIII. On lines and planes of closest fit to systems of points in space , 1901 .