Related papers: A Tale of Two Fields: Neural Network-Enhanced non-…
We present the halo number counts and its two-point statistics, the observable angular power spectrum, extracted for the first time from relativistic N-body simulations. The halo catalogues used in this work are built from the relativistic…
We study structure formation in the presence of primordial non-Gaussianity of the local type with parameters f_NL and g_NL. We show that the distribution of dark-matter halos is naturally described by a multivariate bias scheme where the…
In this paper we study the problem of learning a shallow artificial neural network that best fits a training data set. We study this problem in the over-parameterized regime where the number of observations are fewer than the number of…
Constraining primordial non-Gaussianities (PNG) is one of the main goals of new-generation large-scale galaxy surveys. It had been proposed that unclustered tracers (with bias $b_1=0$) could be optimal for PNG studies, and that these could…
Recently, over-parameterized neural networks have been extensively analyzed in the literature. However, the previous studies cannot satisfactorily explain why fully trained neural networks are successful in practice. In this paper, we…
We present a new method that simultaneously solves for cosmology and galaxy bias on non-linear scales. The method uses the halo model to analytically describe the (non-linear) matter distribution, and the conditional luminosity function…
Understanding the halo-galaxy connection is fundamental in order to improve our knowledge on the nature and properties of dark matter. In this work we build a model that infers the mass of a halo given the positions, velocities, stellar…
We report the first result of weak gravitational lensing survey on a 2.1 sq deg Rc-band image taken with a wide field camera (Suprime-Cam) on the prime focus of 8.2 m Subaru Telescope. The weak lensing mass reconstruction is applied to the…
Conservative mass limits are often imposed on the dark matter halo catalogues extracted from N-body simulations. By comparing simulations with different mass resolutions, at $z=0$ we find that even for halos resolved by 100 particles, the…
We study the non-Gaussianity of the large-scale clustering of high-redshift halos, seeking to assess which terms of standard bias expansions are needed to understand these highly biased populations. We find that the clustering can be well…
Feature selection, as a critical pre-processing step for machine learning, aims at determining representative predictors from a high-dimensional feature space dataset to improve the prediction accuracy. However, the increase in feature…
A novel general neural network (GNN) is proposed for two-class data mining in this study. In a GNN, each attribute in the dataset is treated as a node, with each pair of nodes being connected by an arc. The reliability is of each arc, which…
We study the relationship between dark-matter haloes and matter in the MIP $N$-body simulation ensemble, which allows precision measurements of this relationship, even deeply into voids. What enables this is a lack of discreteness,…
Local primordial non-Gaussianity (LPNG) couples long-wavelength cosmological fluctuations to the short-wavelength behavior of galaxies. This coupling is encoded in bias parameters including $b_{\phi}$ and $b_{\delta\phi}$ at linear and…
We give an analytical form for the weighted correlation function of peaks in a Gaussian random field. In a cosmological context, this approach strictly describes the formation bias and is the main result here. Nevertheless, we show its…
We describe a new algorithm for finding substructures within dark matter haloes from N-body simulations. The algorithm relies upon the fact that dynamically distinct substructures in a halo will have a {\em local} velocity distribution that…
Deep neural networks generalize well despite being heavily overparameterized, in apparent contradiction with classical learning theory based on uniform convergence over fixed hypothesis spaces. Uniform bounds over the entire parameter space…
Context. The spatial distribution of haloes in the Cosmic Web encodes a wealth of information about the underlying cosmological model. These haloes can be represented as nodes of a graph, whose structural properties reflect cosmological…
Large neural networks have proved remarkably effective in modern deep learning practice, even in the overparametrized regime where the number of active parameters is large relative to the sample size. This contradicts the classical…
The structural and dynamic properties of the dark matter halos, though an important ingredient in understanding large-scale structure formation, require more conservative particle resolution than those required by halo mass alone in a…