We report the first observation of the parity-violating gamma-ray asymmetry A np γ in neutronproton capture using polarized cold neutrons incident on a liquid parahydrogen target at the Spallation Neutron Source at Oak Ridge National Laboratory. A np γ isolates the ∆I = 1, 3 S1 → 3 P1 component of the weak nucleon-nucleon interaction, which is dominated by pion exchange and can be directly related to a single coupling constant in either the DDH meson exchange model or pionless effective field theory. We measured A np γ = (−3.0 ± 1.4(stat.) ± 0.2(sys.)) × 10 −8 , which implies a DDH weak πN N coupling of h 1 π = (2.6 ± 1.2(stat.) ± 0.2(sys.)) × 10 −7 and a pionless EFT constant of C 3 S 1 → 3 P 1 /C0 = (−7.4 ± 3.5(stat.) ± 0.5(sys.)) × 10 −11 MeV −1 . We describe the experiment, data analysis, systematic uncertainties, and implications of the result.

We present and discuss a new method to extract parton distribution functions from hard scattering processes based on an alternative type of neural network, the Self-Organizing Map. Quantitative results including a detailed treatment of uncertainties are presented within a Next to Leading Order analysis of inclusive electron proton deep inelastic scattering data.

We discuss the application of an alternative type of neural network, the Self-Organizing Map to extract parton distribution functions from various hard scattering processes. PACS numbers: 13.60.Hb, 13.40.Gp, 24.85.+p I. INTRODUCTIONArtificial Neural Networks (ANNs) are an algorithm model inspired by the human brain's capacity to perform the complex operations of learning, memorizing and generalizing. The goal of ANNs is, however, to solve objective problems which are by far less complex relatively to the human brain capabilities. Their basic units are sets of nodes that are defined as neurons because they can take sets of input parameters and either retain or communicate a signal in a similar way to how signals propagate from one neuron to the other as the neurons get activated/fire. This procedure is defined via learning algorithms. Its main success is in that it allows one to estimate non-linear functions of input data.ANNs consist of a set of initial data forming an input layer, a process by which the input data are evolved and trained (hidden layers), and a resulting set of output data, or the output layer (Figure 1). Researchers currently utilize ANNs in data visualization, function modeling and approximating values of functions, data processing, robotics and control engineering. In the past twenty years ANNs have also established their role as a remarkable computational tool in high energy, nuclear and computational physics analyses. Important applications have been developed, for instance, as classification methods for off-line jet identification and tracking, on-line process control or event trigger and mass reconstruction, and optimization techniques in e.g. track finding and classification [1].Neural networks are used for modeling in these fields of physics because they can utilize the principle of learning with regards to data sets and models. The learning can be supervised or unsupervised.

No abstract

No abstract