Determinant Calculations Using Random Walk Worldline Loops
We use statistical ensembles of worldline loops generated by random walk on hypercubic lattices to calculate matter determinants in background Yang-Mills fields.
arXiv subjects
Publications and source records attributed to Ion-Olimpiu Stamatescu.
We use statistical ensembles of worldline loops generated by random walk on hypercubic lattices to calculate matter determinants in background Yang-Mills fields.
We propose to calculate bosonic and fermionic determinants with some general field background, and the corresponding 1-loop effective actions by evaluating random walk worldline loops generated statistically on the lattice. This is illustrated by some numerical calculations for constant gauge field backgrounds and then discussed for the general case.
We study the κ^2 corrections to the quenched limit of μ>0 QCD. We use an improved reweighting procedure.
We study a simple learning model based on the Hebb rule to cope with "delayed", unspecific reinforcement. In spite of the unspecific nature of the information-feedback, convergence to asymptotically perfect generalization is observed, with a rate depending, however, in a non- universal way on learning parameters. Asymptotic convergence can be as fast as that of Hebbian learning, but may be slower. Moreover, for a certain range of parameter settings, it depends on initial conditions whether the system can reach the regime of asymptotically perfect generalization, or rather approaches a stationary state of poor generalization.
We study some features of learning models based on "delayed" and undifferentiated reinforcement and realized by simple algorithms which may be considered of a very elementary nature. We show that a modification of the Hebb-rule works well for this problem in a neural network realization and study numerically its convergence properties. An illustration for a more "concrete" situation is provided.
We study on the lattice the topology of SU(2) and SU(3) Yang-Mills theories at zero temperature and of QCD at temperatures around the phase transition. To smooth out dislocations and the UV noise we cool the configurations with an action which has scale invariant instanton solutions for instanton size above about 2.3 lattice spacings. The corresponding "improved" topological charge stabilizes at an integer value after few cooling sweeps. At zero temperature the susceptibility calculated from this charge (about (195MeV)^4 for SU(2) and (185 MeV)^4 for SU(3)) agrees very well with the phenomenological expectation. At the minimal amount of cooling necessary to resolve the structure in terms of instantons and anti-instantons we observe a dense ensemble where the total number of peaks is by a factor 5-10 larger than the net charge. The average size observed for these peaks at zero temperature is about 0.4-0.45 fm for SU(2) and 0.5-0.6 fm for SU(3). The size distribution changes very little with further cooling, although in this process up to 90% of the peaks disappear by pair annihilation. For QCD we observe below T_c a reduction of the topological susceptibility as an effect of the dynamical fermions. Nevertheless also here the instantons form a dense ensemble with general characteristics similar to those of the quenched theory. A further drop in the susceptibility above T_c is also in rough agreement with what has been observed for pure SU(3). We see no clear signal for dominant formation of instanton - anti-instanton molecules.
We study the topological structure of the SU(2) vacuum at zero temperature: topological susceptibility, size, shape and distance distributions of the instantons. We use a cooling algorithm based on an improved action with scale invariant instanton solutions. This algorithm needs no monitoring or calibration, has an inherent cut off for dislocations and leaves unchanged instantons at physical scales. The physical relevance of our results is checked by studying the scaling and finite volume dependence. We obtain a susceptibility of (200(15) MeV)^4. The instanton size distribution is peaked around 0.43fm, and the distance distribution indicates a homogeneous, random spatial structure.
We use a cooling algorithm based on an improved action with scale invariant instanton solutions, which needs no monitoring or calibration and has a inherent cut off for dislocations. We present results for SU(2) Yang-Mills theory where the method provides good susceptibility data and physical size distributions of instantons.
We propose and study a ``gold-washing" - type of algorithm which smooths out the short range fluctuations but leaves invariant instantons above a certain size. The algorithm needs no monitoring or calibration.