Search for: All records

In this paper, the neutral 2SC phase of color superconductivity is investigated in the presence of a magnetic field and for diquark coupling constants and baryonic densities that are expected to characterize neutron stars. Specifically, the behavior of the charged gluons Meissner masses is investigated in the parameter region of interest, taking into account, in addition, the contribution of a rotated magnetic field. It is found that up to moderately high diquark coupling constants the mentioned Meissner masses become tachyonic independently of the applied magnetic-field amplitude, hence signalizing the chromomagnetic instability of this phase. To remove the instability, the restructuring of the system ground state is proposed, which now will be formed by vortices of the rotated charged gluons. These vortices boost the applied magnetic field, having the most significant increase for relatively low applied magnetic fields. Finally, considering that with the stellar rotational frequency observed for magnetars a field of the order of 10^8 G can be generated by dynamo effect, we show that by the boosting effect just described the field can be amplified to 10^17 G that is in the range of inner core fields expected for magnetars. Thus, we conclude that the described mechanism could be the one responsible for the large fields characterizing magnetars if the cores of these compact objects are formed by neutral 2SC matter. 

Learning-to-learn (using optimization algorithms to learn a new optimizer) has successfully trained efficient optimizers in practice. This approach relies on meta-gradient descent on a meta-objective based on the trajectory that the optimizer generates. However, there were few theoretical guarantees on how to avoid meta-gradient explosion/vanishing problems, or how to train an optimizer with good generalization performance. In this paper, we study the learning-to-learn approach on a simple problem of tuning the step size for quadratic loss. Our results show that although there is a way to design the meta-objective so that the meta-gradient remain polynomially bounded, computing the meta-gradient directly using backpropagation leads to numerical issues that look similar to gradient explosion/vanishing problems. We also characterize when it is necessary to compute the meta-objective on a separate validation set instead of the original training set. Finally, we verify our results empirically and show that a similar phenomenon appears even for more complicated learned optimizers parametrized by neural networks. 

Abstract Topological spin/polarization structures in ferroic materials continue to draw great attention as a result of their fascinating physical behaviors and promising applications in the field of high‐density nonvolatile memories as well as future energy‐efficient nanoelectronic and spintronic devices. Such developments have been made, in part, based on recent advances in theoretical calculations, the synthesis of high‐quality thin films, and the characterization of their emergent phenomena and exotic phases. Herein, progress over the last decade in the study of topological structures in ferroic thin films and heterostructures is explored, including the observation of topological structures and control of their structures and emergent physical phenomena through epitaxial strain, layer thickness, electric, magnetic fields, etc. First, the evolution of topological spin structures (e.g., magnetic skyrmions) and associated functionalities (e.g., topological Hall effect) in magnetic thin films and heterostructures is discussed. Then, the exotic polar topologies (e.g., domain walls, closure domains, polar vortices, bubble domains, and polar skyrmions) and their emergent physical properties in ferroelectric oxide films and heterostructures are explored. Finally, a brief overview and prospectus of how the field may evolve in the coming years is provided.