Search for: All records

Abstract The non-perturbative Landau-Khalatnikov-Fradkin (LKF) transformations describe how Green functions in quantum field theory transform under a change in the photon field’s linear covariant gauge parameter (denotedξ). The transformations are framed most simply in coordinate space where they are multiplicative. They imply that information on gauge-dependent contributions from higher order diagrams in the perturbative series is contained in lower order contributions, which is useful in multi-loop calculations. We study the LKF transformations for the propagator and the vertex in both scalar and spinor QED, in some particular dimensions. A novelty of our work is to derive momentum-space integral representations of these transformations; our expressions are also applicable to the longitudinal and transverse parts of the vertex. Applying these transformations to the tree-level Green functions, we show that the one-loop terms obtained from the LKF transformation agree with the gauge dependent parts obtained from perturbation theory. Our results will be presented in more comprehensive form elsewhere. 

In this work we present DRUM, an unsupervised approach that is based on statistical properties of multivariate data streams to identify regime shifts in real time. DRUM processes streams in small chunks, learns their statistical properties, and makes generalizations as time goes by. We show how this straightforward approach requires minimal computation and reaches state of the art accuracy, making it ideal for embedded and cyber physical systems. 

This dataset contains the electric power consumption data from the Los Alamos Public Utility Department (LADPU) in New Mexico, USA. The data was collected by Landis+Gyr smart meters devices on 1,757 households at North Mesa, Los Alamos, NM. The sampling rate is one observation every fifteen minutes (i.e., 96 observations per day). For most customers, the data spans about six years, from July 30, 2013 to December 30, 2019. However, for some customers, the period is reduced. The dataset contains missing values and duplicated measurements.  This dataset is provided in its original format, without cleaning or pre-processing. The only procedure performed was for anonymization reasons. Thus, the data are not normalized, and it has missing values and duplicate entries (i.e., more than one measurement for the same time). However, these issues represent only a small portion of data.