New information demonstrating the importance of both sequential and simultaneous (or direct) multiphoton ionization of inner shell electrons from neon is discussed in this paper. Ne was irradiated with intense 93 eV free electron laser (FEL) pulses at FLASH and studied with the aid of photoelectron spectrometry. This resulted in two and three photon, single and double ionization of neon, removing electrons from 2s and 2p subshells of the neutral Ne atom in multiple different pathways. The spectral features of the photoelectrons were identified through comparison with the NIST database and field averaged time-dependent density matrix theory. The calculations show the direct multiphoton ionization processes to be extremely sensitive to the focused FEL intensity.

MP-CITDSE is a package of programs which solves the time-dependent Schrodinger equation for hydrogenic and helium-like atomic systems interacting with an ultra-short laser pulse (of attosecond or femtosecond duration). The output of the computations - with some minimal processing - may be used to calculate excited populations, single and double ionisation yields, kinetic, angular and radial electron distributions, and harmonic yields. For the Helium atom, a full account of the inter-electronic correlation effects is included via a configuration interaction approach; for the time propagation of the wavefunction a spectral basis expansion on the eigenstates of the field-free Hamiltonian is used; for this reason post -processing of the expansion coefficients just after the pulse lead to simple formulas for quantities of experimental interest. Program summary Program Title: MP-CITDSE CPC Library link to program files: https://doi .org /10 .17632 /gv4zxmfdx3 .1 Developer's repository link: https://github .com /aforembs /MP-CITDSE Licensing provisions: GPLv3 Programming language: C++ Nature of problem: The ultra-fast high-intensity laser pulses of modern experimental sources require an ab-initio theoretical treatment, as the characteristics of such pulses are generally incompatible with the established Lowest-order Perturbation Theory (LOPT) approximation. Such a treatment can be quite cumbersome to not only implement, but also describe in a clear, intuitive manner. Its inherent complexity ensures that even in its simplest form, an ab-initio TDSE solver can easily consist of several thousand lines of code. As with any project of such scope a significant amount of effort is required to keep the code up to date, running on several platforms, compilers and operating systems. As such, while the numerical problem tackled by MP-CITDSE is that of the simulation of hydrogenic and helium-like atom-laser interactions on a single node computer/workstation; its main challenges are longevity and expandability. By doing away with unnecessary abstractions within the code, using a modern standard of a widely used and performant language (C++ 17) and hosting the source code on a publicly available GitHub repository we hope that the programs provided herein prove themselves useful to many current and future AMO Physicists. Solution method: The atomic systems are considered confined in a spherical box of finite size. In the one-electron case (hydrogen) the eigenstates are calculated in a spherical coordinate system by the direct diagonalisation of the field-free Hamiltonian matrix representation on a B-splines piecewise polynomial basis expansion of the radial part of the eigenfunctions; for the two-electron case (helium) the numerically calculated one-electron eigenstates are angularly coupled to form a zero-order (uncorrelated) two-electron numerical basis; then the latter basis is further coupled via a configuration-interaction approach to eventually calculate the helium's two-electron eigenstates. Following these steps, for both systems (hydrogen and helium) the computation proceeds by expanding the time-dependent wavefunction on the corresponding field-free eigenstate basis; thus, we end up to a system of first-order ordinary differential equations (ODEs) in time for the expansion (time-dependent) coefficients; the dynamical parameters for the ODEs time propagation, namely the eigenenergies and dipole matrix elements are calculated only once prior to the ODEs propagation [2]. The structure of the propagating matrix is block tridiagonal. Additional comments including restrictions and unusual features: This paper serves as the definitive reference for the MP-CITDSE code. In this version we solve the ODEs using a Runge-Kutta-Felnberg (RKF) algorithm. Also we have restricted the code to treat hydrogen and helium but with very little effort (non-relativistic) hydrogenic and helium-like atomic systems can also be simulated. The programs are supplied with two methods of compilation, via make and cmake to ensure greater portability.

Two XUV-photon double ionization of Ne, induced by an intense few-pulse attosecond train with a 4 fs envelope duration is investigated experimentally and theoretically. The experiment is performed at ELI-ALPS utilizing the recently constructed 10 Hz gas phase high-order harmonic generation SYLOS GHHG-COMPACT beamline. A total pulse energy up to 1 μJ generated in Argon in conjunction with high reflectivity optics in the XUV region, allowed the observation of the doubly charged state of Ne induced by 40 eV central XUV photon energies. The interaction of the intense attosecond pulse train with Ne is also theoretically studied via a second-order time dependent perturbation theory equations-of-motion. The results of this work, combined with the feasibility of conducting XUV-pump-XUV-probe experiments, constitute a powerful tool for many potential applications. Those include attosecond pulse metrology as well as time resolved investigations of the dynamics underlying direct and sequential double ionization and their electron correlation effects.

Two-XUV-photon double ionization of Ne, induced by an intense few-pulse attosecond train with a -4 fs envelope duration is investigated experimentally and theoretically. The experiment is performed at ELI-ALPS (Extreme Light Infrastructure Attosecond Light Pulse Source) utilizing the recently constructed 10 Hz gas phase high-order harmonic generation SYLOS GHHG-COMPACT beamline. A total pulse energy up to -1 mu J generated in argon in conjunction with high-reflectivity optics in the XUV region allowed the observation of the doubly charged state of Ne induced by 40 eV central XUV-photon energies. The interaction of the intense attosecond pulse train with Ne is also theoretically studied via second-order time-dependent perturbation theory equations of motion. The results of this work, combined with the feasibility of conducting XUV-pump-XUV-probe experiments, constitute a powerful tool for many potential applications. Those include attosecond pulse metrology as well as time-resolved investigations of the dynamics underlying direct and sequential double ionization and their electron correlation effects.

Motivated by the achieved high intensities of novel extreme ultraviolet (XUV) radiation sources, such as free electron lasers and laser-driven high harmonic generation beamlines, we elaborate on their perspective in inducing observable strong field effects. The feasibility of extending such effects from the infrared and visible spectral regimes in the XUV domain is supported through numerically calculated models of near-future experiments. We highlight the advancement of performing studies in the time domain, using ultra-short XUV pulses, which allows for the temporal evolution of such effects to be followed. Experimental and theoretical obstacles and limitations are further discussed.

2D Convolutional Neural Networks (CNNs) have enjoyed a surge in popularity over the last few years, mainly because they outperform traditional algorithms/methods in a myriad of computer vision (and other fields) tasks. On the other hand, the problem becomes more complex when dealing with 3D volumes. Lack of readily available training data, memory and computational requirements are just some of the factors hindering the progress of 3D CNNs. We propose a synthetic 3D voxelized point-clouds generation method containing object and scene in this paper. Furthermore, an efficient 3D volumetric representation called VOLA is applied. VOLA (Volumetric Accelerator) is a sexaquaternary (power-of-four subdivision) tree based representation which aims to save significant memory for volumetric data. After training the model, it was deployed onto Movidius Neural Compute Stick which is a USB, containing a low-power processing unit as well as dedicated CNN hardware blocks. The trained model on NCS takes only 90 frames per second to perform inference on each 3D volume, with an average power consumption of 1.2W.

This article presents the first known benchmark of Convolutional Neural Networks (CNN) with a focus on inference time and power consumption. This benchmark is important for low-cost and low-power robots running on batteries where it is required to obtain good performance at the minimal power consumption. The CNN are benchmarked either on a Raspberry Pi, Intel Joule 570X or on a USB Neural Compute Stick (NCS) based on the Movidius MA2450 Vision Processing Unit (VPU). Inference performance is evaluated for TensorFlow and Caffe running either natively on Raspberry Pi or Intel Joule and compared against hardware acceleration obtained by using the NCS. The use of the NCS demonstrates up to 4x faster inference time for some of the networks while keeping a reduced power consumption. For robotics applications this translates into lower latency and smoother control of a robot with longer battery life and corresponding autonomy. As an application example, a low-cost robot is used to follow a target based on the inference obtained from a CNN.