Towards reduction of autocorrelation in HMC by machine learning

In this paper we propose new algorithm to reduce autocorrelation in Markov chain Monte-Carlo algorithms for euclidean field theories on the lattice. Our proposing algorithm is the Hybrid Monte-Carlo algorithm (HMC) with restricted Boltzmann machine. We examine the validity of the algorithm by employing the phi-fourth theory in three dimension. We observe reduction of the autocorrelation both in symmetric and broken phase as well. Our proposing algorithm provides consistent central values of expectation values of the action density and one-point Green's function with ones from the original HMC in both the symmetric phase and broken phase within the statistical error. On the other hand, two-point Green's functions have slight difference between one calculated by the HMC and one by our proposing algorithm in the symmetric phase. Furthermore, near the criticality, the distribution of the one-point Green's function differs from the one from HMC. We discuss the origin of discrepancies and its improvement.

[1]  Testing algorithms for critical slowing down , 2018, 1710.07036.

[2]  Roger G. Melko,et al.  Deep Learning the Ising Model Near Criticality , 2017, J. Mach. Learn. Res..

[3]  Léon Bottou,et al.  Wasserstein Generative Adversarial Networks , 2017, ICML.

[4]  M. Hasenbusch.  Fighting topological freezing in the two-dimensional CPN-1 model , 2017, 1709.09460.

[5]  Yang Qi,et al.  Self-learning Monte Carlo method: Continuous-time algorithm , 2017, 1705.06724.

[6]  Yang Qi,et al.  Self-Learning Monte Carlo Method in Fermion Systems , 2016, 1611.09364.

[7]  Yang Qi,et al.  Self-learning Monte Carlo method , 2016, 1610.03137.

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

[9]  Xiao Yan Xu,et al.  Self-learning quantum Monte Carlo method in interacting fermion systems , 2016, 1612.03804.

[10]  Geoffrey E. Hinton,et al.  Deep Learning , 2015, Nature.

[11]  Yoshua Bengio,et al.  Generative Adversarial Nets , 2014, NIPS.

[12]  Frithjof Karsch,et al.  Thermodynamics of strong-interaction matter from Lattice QCD , 2013, 1504.05274.

[13]  Geoffrey E. Hinton.  A Practical Guide to Training Restricted Boltzmann Machines , 2012, Neural Networks: Tricks of the Trade.

[14]  Tapani Raiko,et al.  Improved Learning of Gaussian-Bernoulli Restricted Boltzmann Machines , 2011, ICANN.

[15]  S. Schaefer,et al.  Critical slowing down and error analysis in lattice QCD simulations , 2010, 1009.5228.

[16]  T. Hatsuda,et al.  Theoretical Foundation of the Nuclear Force in QCD and Its Applications to Central and Tensor Forces in Quenched Lattice QCD Simulations , 2009, 0909.5585.

[17]  S. Schaefer,et al.  Investigating the critical slowing down of QCD simulations , 2009, 0910.1465.

[18]  Yoshua Bengio,et al.  Justifying and Generalizing Contrastive Divergence , 2009, Neural Computation.

[19]  P. Forcrand,et al.  Testing and tuning new symplectic integrators for Hybrid Monte Carlo algorithm in lattice QCD , 2008 .

[20]  S. Aoki,et al.  Two-flavor lattice QCD in the epsilon-regime and chiral Random Matrix Theory , 2007, 0705.3322.

[21]  Yee Whye Teh,et al.  A Fast Learning Algorithm for Deep Belief Nets , 2006, Neural Computation.

[22]  Martin Luscher.  Schwarz-preconditioned HMC algorithm for two-flavour lattice QCD , 2004 .

[23]  U. Wolff,et al.  Monte Carlo errors with less errors , 2003, hep-lat/0306017.

[24]  K. Jansen.  Actions for dynamical fermion simulations: are we ready to go? , 2003, hep-lat/0311039.

[25]  G. Hooft.  Confinement of quarks , 2003 .

[26]  Geoffrey E. Hinton.  Training Products of Experts by Minimizing Contrastive Divergence , 2002, Neural Computation.

[27]  M. Hasenbusch.  Speeding up the hybrid Monte Carlo algorithm for dynamical fermions , 2001, hep-lat/0110180.

[28]  M. Burkardt.  Gauge field theories on a ⊥ lattice , 1999, hep-th/9908195.

[29]  T. Blum.  The QCD Equation of State from the Lattice. , 1996 .

[30]  J. Sexton,et al.  Hamiltonian evolution for the hybrid Monte Carlo algorithm , 1992 .

[31]  A. Sokal,et al.  The pivot algorithm: A highly efficient Monte Carlo method for the self-avoiding walk , 1988 .

[32]  A. Kennedy,et al.  Hybrid Monte Carlo , 1988 .

[33]  M. Lüscher.  Construction of a selfadjoint, strictly positive transfer matrix for Euclidean lattice gauge theories , 1977 .

[34]  S. Elitzur,et al.  Impossibility of spontaneously breaking local symmetries , 1975 .