Cosmological simulations provide a wealth of data in the form of point clouds and directed trees. A crucial goal is to extract insights from this data that shed light on the nature and composition of the Universe. In this paper we introduce CosmoBench, a benchmark dataset curated from state-of-the-art cosmological simulations whose runs required more than 41 million core-hours and generated over two petabytes of data. CosmoBench is the largest dataset of its kind: it contains 34 thousand point clouds from simulations of dark matter halos and galaxies at three different length scales, as well as 25 thousand directed trees that record the formation history of halos on two different time scales. The data in CosmoBench can be used for multiple tasks -- to predict cosmological parameters from point clouds and merger trees, to predict the velocities of individual halos and galaxies from their collective positions, and to reconstruct merger trees on finer time scales from those on coarser time scales. We provide several baselines on these tasks, some based on established approaches from cosmological modeling and others rooted in machine learning. For the latter, we study different approaches -- from simple linear models that are minimally constrained by symmetries to much larger and more computationally-demanding models in deep learning, such as graph neural networks. We find that least-squares fits with a handful of invariant features sometimes outperform deep architectures with many more parameters and far longer training times. Still there remains tremendous potential to improve these baselines by combining machine learning and cosmology to fully exploit the data. CosmoBench sets the stage for bridging cosmology and geometric deep learning at scale. We invite the community to push the frontier of scientific discovery by engaging with this dataset, available at this https URL
Understanding how the dynamical state of the interstellar medium (ISM) changes across spatial scales can provide important insights into how the gas is organized and ultimately collapses to form stars. To this end, we present ALMA $^{12}\mathrm{CO}(2-1)$ observations at $7$ pc ($0.4''$) spatial resolution across a $1.4~\mathrm{kpc}\times5.6~\mathrm{kpc}$ ($1'.3\times1'.3$) region located in the disk of the nearby ($D = 3.5$ Mpc), massive, star-forming galaxy NGC 253. We decompose this emission with a hierarchical, multiscale dendrogram algorithm to identify 2463 structures with deconvolved sizes ranging from $\sim3$ to $300$ pc, complete to a limiting mass of $10^4~\mathrm{M_\odot}$. By comparing the virial parameter of these structures against physical properties including size, mass, surface density, velocity dispersion, and hierarchical position, we carry out a comprehensive search for a preferred scale at which gravitationally bound structures emerge. Ultimately, we do not identify any emergent scale for bound objects in our data, nor do we find a correlation between the virial parameter and structure sizes. These findings suggest that gravitational binding cannot be used to define molecular clouds and emphasize the need for multiscale approaches to characterize the ISM.
The China Space Station Telescope (CSST) is a next-generation Stage-IV sky survey telescope, distinguished by its large field of view (FoV), high image quality, and multi-band observation capabilities. It can simultaneously conduct precise measurements of the Universe by performing multi-color photometric imaging and slitless spectroscopic surveys. The CSST is equipped with five scientific instruments, i.e. Multi-band Imaging and Slitless Spectroscopy Survey Camera (SC), Multi-Channel Imager (MCI), Integral Field Spectrograph (IFS), Cool Planet Imaging Coronagraph (CPI-C), and THz Spectrometer (TS). Using these instruments, the CSST is expected to make significant contributions and discoveries across various astronomical fields, including cosmology, galaxy and active galactic nuclei (AGN), the Milky Way and nearby galaxies, stars, exoplanets, Solar System objects, astrometry, and transients and variable sources. This review aims to provide a comprehensive overview of the CSST instruments, observational capabilities, data products, and scientific potential.
Juno and Cassini have shown that Jupiter and Saturn likely contain extended gradients of heavy elements. Yet, how these gradients can survive over billions of years remains an open question. Classical convection theories predict rapid mixing and homogenization, which would erase such gradients on timescales far shorter than the planets' ages. To address this, we estimate the energy required to erode both dense and fuzzy cores, and compare it to what the planet can realistically supply. If the entire cooling budget is available to drive mixing, then even a compact core can, in principle, be destroyed. But if mixing is limited to the thermal energy near the core, which is another plausible scenario, the energy falls short. In that case, Jupiter can erode a fuzzy core by up to approximately $10~\mearth$, but a compact one remains intact. Saturn's core is more robust. Even in the fuzzy case, only about $1~\mearth$ is lost, and if the core is compact, erosion is negligible. The outcome depends sensitively on the assumed initial temperature and entropy profiles. Hotter and more superadiabatic interiors are more prone to mixing. We suggest that 3D simulations of convection driven from above, with realistic stratification and enough depth (i.e., many density scale heights) would be of great interest to further constrain the energy budget for core erosion.
Hot Jupiters (HJs) are commonly thought to host the strongest dynamo-generated magnetic fields among exoplanets, up to one order of magnitude larger than Jupiter. Thus, they have often been regarded as the most promising exoplanets to display magnetic star-planet interaction signals and magnetically-driven coherent radio emission, which unfortunately remains elusive, despite many diversified observational campaigns. In this work, we investigate the evolution of the internal convection and dynamo properties of HJs via one-dimensional models. We explore the dependency on orbital distance, planetary and stellar masses, and types of heat injection. We employ one-dimensional evolutionary models to obtain internal convective structures. Specifically, we obtain the Rossby number $\mathrm{Ro}$ as a function of planetary depth and orbital period, after showing that tidal synchronization is likely valid for all HJs. When the heat is applied uniformly, the convective layers of almost all HJs remain in the fast rotator regime, $\mathrm{Ro} \lesssim 0.1$, except possibly the most massive planets with large orbital distances (but still tidally locked). We recover magnetic field strengths for inflated HJs by applying well-known scaling laws for fast rotators. When strong heat sources are applied mostly in the outer envelope and outside the dynamo region, as realistic Ohmic models predict, convection in the dynamo region often breaks down. Consequently, the heat flux and the derived surface magnetic fields can be greatly reduced to or below Jovian values, contrary to what is commonly assumed, thus negatively affecting estimates for coherent radio emission, and possibly explaining the failure in detecting it so far.
this http URL @oeaw. this http URL