We present a dynamic network rewiring (DNR) method to generate pruned deep neural (DNN) models that both are robust against adversarially generated images and maintain high accuracy on clean images. In particular, the disclosed DNR training is based unified constrained optimization formulation using novel hybrid loss function merges sparse learning with adversarial training. This strategy dynam...

We present a numerical kinetic scheme for an unsteady mixed pressurized and free surface model. This model has a source term depending on both the space variable and the unknown U of the system. Using the Finite Volume and Kinetic (FVK) framework, we propose an approximation of the source terms following the principle of interfacial upwind with a kinetic interpretation. Then, several numerical ...

We present a Higgsless Standard Model in six dimensions, based on the Standard Model gauge group SU(2)L × U(1)Y , with two flat extra dimensions compactified on a rectangle. The electroweak symmetry is broken by (mixed) boundary conditions and realistic gauge boson masses can be accommodated by proper choice of the compactification scales and brane kinetic terms. With respect to “oblique” corre...

We review the construction and theoretical implications of the USp(2k) matrix model in zero dimension introduced in ref. 1), . It is argued that the model provides a constructive approach to TypeI superstrings and is at the same time dynamical theory of spacetime points. Three subjects are discussed : semiclassical pictures and series of degenerate perturbative vacua associated with the worldvo...

We propose U -statistics-based tests for null variance components in linear mixed models. We obtain their asymptotic distribution (for increasing number of units) under mild regularity conditions that include only the existence of the second moment for the distribution of the random effects and of the fourth moment for the distribution of the conditional errors. We also derive their nonnull dis...

Mixed linear models can be used to improve the informational value of milk yield forecasts. For this purpose two different functional approaches for modelling lactation curves are as well compared as three linear mixed models with varying random effects of individual animals and lactation numbers. It can be shown that more complex random regression models fit significantly better than fixed reg...

Transformation of the response of a linear model is a popular method in practice when attempting to satisfy the assumptions of the model. Environmental research routinely uses log-transformations due to the nature of the observed data. The choice of the transformation is often made based upon previous experience or on the comparison of models with different transformed responses. Often a transf...

In our talk, we propose an adaptive mesh-refining strategy for the cell-centered FVM based on some a posteriori error control for the quantity ‖∇T (u − Iuh)‖L2 . Here, uh ∈ P (T ) denotes the FVM approximation of u and I is a certain interpolation operator. As model example serves the Laplace equation with mixed boundary conditions, where our contributions extend a result of [NIC05]. Moreover, ...

Using the Fujikawa and the heat-kernel methods we make a complete and detailed computation of the global, gauge, and gravitational anomalies in the standard model defined on a curved space-time with torsion. We find new contributions coming from the curvature and the torsion to the leptonic number anomaly ~so that B-L is not conserved anymore! to the U~1!Y gauge and to the mixed U~1!Y-gravitati...