Abstract: Predicting coexistence patterns is a current challenge to understand diversity maintenance, especially in rich communities where these patterns' complexity is magnified through indirect interactions that prevent their approximation with classical experimental approaches. We explore cutting-edge Machine Learning techniques called Generative Artificial Intelligence (GenAI) to predict species coexistence patterns in vegetation patches, training generative adversarial networks (GAN) and variational AutoEncoders (VAE) that are then used to unravel some of the mechanisms behind community assemblage. The GAN accurately reproduces real patches' species composition and plant species' affinity to different soil types, and the VAE also reaches a high level of accuracy, above 99%. Using the artificially generated patches, we found that high-order interactions tend to suppress the positive effects of low-order interactions. Finally, by reconstructing successional trajectories, we could identify the pioneer species with larger potential to generate a high diversity of distinct patches in terms of species composition. Understanding the complexity of species coexistence patterns in diverse ecological communities requires new approaches beyond heuristic rules. Generative Artificial Intelligence can be a powerful tool to this end as it allows to overcome the inherent dimensionality of this challenge.

Abstract: Kohn-Sham density functional calculations are reported for the structures of clusters consisting of a carbonyl sulfide (OCS) molecule with N = 1, 8, 18, and 40 attached He-3 atoms. The N = 1 cluster ground state is highly localized at the molecular waist (donut ring position), but for higher levels of excitation becomes increasingly delocalized. The first magic cluster with 8 atoms has a significant density at both ends of the molecule in addition to the donut ring. With N = 18 He-3 atoms the molecule is enclosed by a magic number closed shell. Another magic stable structure consisting of two nearly isotropically spherical closed shells is found at N = 40. A comparison with calculations for the same sized He-4 clusters show some important similarities, e. g., pile up at the donut ring position but altogether a more diffuse, less anisotropic structure. These results are discussed in the light of the recently analyzed infrared spectra measured in large pure He-3 droplets (N approximate to 1.2 x 10(4)) [B. Sartakov, J. P. Toennies, and A. F. Vilesov, J. Chem. Phys. 136, 134316 (2012)]. The moments of inertia of the 11 atom spherical shell structure, which is consistent with the experimental spectrum, lies between the predicted moments of inertia for N = 8 and N = 18 clusters. Overall the calculations reveal that the structures and energies of small doped He-3 are only slightly more diffuse and less energetic than the same He-4 clusters.

Abstract: We present a detailed description of the electromagnetic filter for the PTOLEMY project to directly detect the Cosmic Neutrino Background (CNB). Starting with an initial estimate for the orbital magnetic moment, the higher-order drift process of E x B is configured to balance the gradient-B drift motion of the electron in such a way as to guide the trajectory into the standing voltage potential along the mid-plane of the filter. As a function of drift distance along the length of the filter, the filter zooms in with exponentially increasing precision on the transverse velocity component of the electron kinetic energy. This yields a linear dimension for the total filter length that is exceptionally compact compared to previous techniques for electromagnetic filtering. The parallel velocity component of the electron kinetic energy oscillates in an electrostatic harmonic trap as the electron drifts along the length of the filter. An analysis of the phase-space volume conservation validates the expected behavior of the filter from the adiabatic invariance of the orbital magnetic moment and energy conservation following Liouville's theorem for Hamiltonian systems. (C) 2019 Elsevier B.V. All rights reserved.