Search

Machine learning techniques exhibit significant performance in discriminating different phases of matter and provide a new avenue for studying phase transitions. We investigate the phase transitions of three dimensional $q$-state Potts model on cubic lattice by using a transfer learning approach, Domain Adversarial Neural Network (DANN). With the unique neural network architecture, it could evaluate the high-temperature (disordered) and low-temperature (ordered) phases, and identify the first and second order phase transitions. Meanwhile, by training the DANN with a few labeled configurations, the critical points for $q=2,3,4$ and $5$ can be predicted with high accuracy, which are consistent with those of the Monte Carlo simulations. These findings would promote us to learn and explore the properties of phase transitions in high-dimensional systems., Comment: 18pages,36 figures

A transfer learning method, Domain Adversarial Neural Network (DANN), is introduced to study the phase transition of two-dimensional q-state Potts model. With the DANN, we only need to choose a few labeled configurations automatically as input data, then the critical points can be obtained after training the algorithm. By an additional iterative process, the critical points can be captured to comparable accuracy to Monte Carlo simulations as we demonstrate it for q = 3, 4, 5, 7 and 10. The type of phase transition (first or second-order) is also determined at the same time. Meanwhile, for the second-order phase transition at q=3, we can calculate the critical exponent $\nu$ by data collapse. Furthermore, compared to the traditional supervised learning, we found the DANN to be more accurate with lower cost., Comment: 30 pages, 36 figures

The latest advances of statistical physics have shown remarkable performance of machine learning in identifying phase transitions. In this paper, we apply domain adversarial neural network (DANN) based on transfer learning to studying non-equilibrium and equilibrium phase transition models, which are percolation model and directed percolation (DP) model, respectively. With the DANN, only a small fraction of input configurations (2d images) needs to be labeled, which is automatically chosen, in order to capture the critical point. To learn the DP model, the method is refined by an iterative procedure in determining the critical point, which is a prerequisite for the data collapse in calculating the critical exponent $\nu_{\perp}$. We then apply the DANN to a two-dimensional site percolation with configurations filtered to include only the largest cluster which may contain the information related to the order parameter. The DANN learning of both models yields reliable results which are comparable to the ones from Monte Carlo simulations. Our study also shows that the DANN can achieve quite high accuracy at much lower cost, compared to the supervised learning., Comment: 12 pages,10 figures

We propose a forward method, based on the PYTHIA6.4, to study theoretically the jet properties in the ultra-relativistic pp collisions. In the forward method, the partonic initial states are first generated with PYTHIA6.4 and then hadronized in the Lund srting fragmentation regime, and finally hadronic jets are constructed from the created hadrons. Jet properties calculated in the forward method for pp collisions at \sqrt{s}=7 TeV are comparable with the corresponding ones calculated with usual anti-k_t algorithm (backward method) in the PYTHIA6.4. The comparison between the results in the backward and forward methods may bring benefit to the understanding of the partonic origin of jets in the backward method., Comment: 6 pages, 5 figures

Many models and real complex systems possess critical thresholds at which the systems shift from one sate to another. The discovery of the early warnings of the systems in the vicinity of critical point are of great importance to estimate how far a system is from a critical threshold. Multifractal Detrended Fluctuation analysis (MF-DFA) and visibility graph method have been employed to investigate the fluctuation and geometrical structures of magnetization time series of two-dimensional Ising model around critical point. The Hurst exponent has been confirmed to be a good indicator of phase transition. Increase of the multifractality of the time series have been observed from generalized Hurst exponents and singularity spectrum. Both Long-term correlation and broad probability density function are identified to be the sources of multifractality of time series near critical regime. Heterogeneous nature of the networks constructed from magnetization time series have validated the fractal properties of magnetization time series from complex network perspective. Evolution of the topology quantities such as clustering coefficient, average degree, average shortest path length, density, assortativity and heterogeneity serve as early warnings of phase transition. Those methods and results can provide new insights about analysis of phase transition problems and can be used as early warnings for various complex systems., Comment: 23 pages, 11 figures

The factorial moments analyses are performed to study the scaling properties of the dynamical fluctuations of contacts and nodes in temporal networks based on empirical data sets. The intermittent behaviors are observed in the fluctuations for all orders of the moments. It indicates that the interaction has self-similarity structure in time interval and the fluctuations are not purely random but dynamical and correlated. The scaling exponents for contacts in Prostitution data and nodes in Conference data are very close to that for 2D Ising model undergoing a second-order phase transition., Comment: 4 pages, 6 figures

Within the framework of Ginzburg-Landau theory, the effect of multiplicity correlation between the dynamical multiplicity fluctuations is analyzed for a first-order phase transition from quark-gluon plasma to hadrons. Normalized factorial correlators are used to study the correlated dynamical fluctuations. A scaling behavior is found among the factorial correlators, and an approximate universal exponent, which is weakly dependent on the details of the phase transition, is obtained., Comment: 4 pages with 5 eps figures

The features of magnetic field in relativistic heavy-ion collisions are systematically studied by using a modified magnetic field model in this paper. The features of magnetic field distributions in the central point are studied in the RHIC and LHC energy regions. We also predict the feature of magnetic fields at LHC $\sqrt{s_{NN}}$= 900, 2760 and 7000 GeV based on the detailed study at RHIC $\sqrt{s_{NN}}$ = 62.4, 130 and 200 GeV. The dependencies of the features of magnetic fields on the collision energies, centralities and collision time are systematically investigated, respectively., Comment: 8 pages, 7 figures