arXiv subjects

Kana Moriwaki

Publications and source records attributed to Kana Moriwaki.

At least 19 recordsLinked to original sources

arXiv2026

The power spectrum of galaxies from large to small scales: a line-intensity mapping perspective

We present a model for the power spectrum of the density field of galaxies weighted by their star formation rate. This weighting is relevant in line-intensity mapping (LIM) when the observed line luminosity is strongly correlated with star formation, as is the case for the H$α$ line. Our model reproduces the measured power spectrum in the IllustrisTNG simulation to within a few per cent across all scales, with fitting parameters that have clear physical interpretations. On scales of tens of megaparsecs, the model accounts for the weighted non-linear bias of galaxies as well as halo exclusion (2-halo term). On smaller scales, it incorporates the weighted distribution of satellite galaxies within haloes (1-halo term). The random sampling of satellite galaxies introduces a galaxy shot noise term to the power spectrum on small scales, and their confinement to haloes introduces a halo shot noise term on large scales. Omitting satellite galaxies from the analysis results in an underestimation of both the large-scale bias and the mean intensity by approximately 30 per cent each at redshift 1.5. Assigning the intensity of satellites to the centre of their respective haloes affects the power spectrum on scales $k > 0.3$ h Mpc$^{-1}$. Our fitting function provides a well-motivated parametrisation that can be used to interpret data from upcoming LIM surveys.

astro-ph.GA↗

arXiv2026

From Dark Matter to Galaxies: Halo-Free Mock Generation via Conditional Point-Cloud Diffusion

We present a diffusion-based generative model for constructing realistic galaxy catalogues from dark matter density fields. The model takes a three-dimensional dark matter density field as input and generates galaxies directly as a point cloud with positions and physical properties, including star formation rate (SFR), without identifying dark matter haloes or subhaloes. We train the model on galaxy catalogues from the IllustrisTNG hydrodynamical simulation. The generated catalogues reproduce the spatial correspondence between galaxies and the underlying dark matter field, preferentially populating dense regions and filamentary structures. They also accurately reproduce the one- and two-dimensional distributions of SFR and stellar mass, the galaxy auto-power spectrum, and the galaxy--dark matter cross-power spectrum. The model generates galaxies associated with structures below the nominal resolution of the input density field by marginalising over unresolved small-scale structure rather than relying on a resolved halo catalogue. With an optimised diffusion sampling schedule, it generates a catalogue with SFR $> 1 ~\rm M_\odot/yr$ over a $(151.3 ~\rm Mpc)^3$ volume in approximately 10 seconds on a single GPU. Our model therefore provides a practical engine for producing large mock ensembles for upcoming galaxy redshift surveys and line-intensity mapping experiments, and offers a path toward simulation-based inference that bypasses halo finding and directly connects field-level dark matter statistics to observable galaxy populations.

astro-ph.GA↗

arXiv2026

Square Kilometer Array Synergies for the Epoch of Reionization and Cosmic Dawn

Synergies with other instruments will be essential in making, verifying, and interpreting a detection of the cosmic 21-cm signal from the Epoch of Reionization (EoR) and Cosmic Dawn (CD) with the Square Kilometer Array (SKA) telescope. Such synergies can (i) provide prior information about galaxies and the intergalactic medium (IGM) during the EoR/CD; (ii) pave the road to a first 21cm detection by mitigating foregrounds and systematics through cross-correlations; and (iii) give complimentary physical insights into the galaxy -- IGM connection. Here we review the current state of synergies and discuss what observations will best compliment SKA-low EoR/CD observations.

astro-ph.CO↗

arXiv2026

Development of TIFUUN: Terahertz Integral Field Units with Universal Nanotechnology

TIFUUN (THz Integral Field Units with Universal Nanotechnology) is an ultra-wideband mm-submm wave imaging spectrometer that capitalizes on the highly scalable integrated superconducting spectrometer technology. TIFUUN has two slots for integral field units (IFUs), which can jointly be optimized as open-hardware for each astronomical observation in terms of spatial and spectral coverage. These IFUs can have observation frequencies in the range of 90--360 GHz, with spectral resolution up to $R\equiv F/ΔF \le 1,000$, with up to $\sim$18,000 kinetic inductance detectors (shared by the two IFUs with a flexible ratio). The ultra-wide 4:1 (2 octave) bandwidth optics fits in a remarkably compact volume, by means of thin silicon lenses and a high chief ray angle design. The first pair of IFUs are being developed for the SUBLIME (Study of the Universe By Line Intensity Mapping Experiments) experiment that aims to map CII emission at redshift $\sim$6 to trace the cosmic large-scale structure and the buildup of galaxies during reionization, using TIFUUN on the ASTE 10-m telescope. The scalability, flexibility and compactness makes TIFUUN a highly compatible and portable system suited also for upcoming telescope facilities in the vicinity, such as FYST and AtLAST/LST.

astro-ph.IM↗

arXiv2026

Signatures of Accreting Black Holes in Line Intensity Mapping

Line-intensity mapping (LIM) has attracted growing attention as a powerful technique for probing the large-scale distribution of galaxies and the cosmic history of star formation through unresolved line emission. Existing LIM models for galaxy-associated lines, such as H$α$, often assume that the dominant contribution to observed emission arises from star-forming activity, while the role of accreting black holes (BHs) remains largely unexplored. In this study, we use the IllustrisTNG cosmological hydrodynamical simulation to construct mock intensity maps of H$α$ and He II, including contributions from both star formation and BH accretion. We show that the BH contribution to the mean intensity is significant, reaching $\sim$40--60 per cent for H$α$ and $\sim$60--80 per cent for He II around cosmic noon. Owing to the large luminosity weight of rare, bright sources, BH-powered emission dominates the shot-noise component of the power spectrum and significantly boosts the small-scale clustering amplitude, particularly for He II. We assess the implications for forthcoming LIM surveys and show that SPHEREx can probe the BH-influenced bright end of the H$α$ voxel intensity distribution (VID) at $z\lesssim4$, and a CDIM-like experiment can further access the BH-dominated regime of He II. Our results demonstrate that accreting BHs represent an essential component of LIM signals, which was previously underappreciated. We thus conclude that accurately modeling the BH contribution is crucial for a physically complete interpretation of future LIM observations.

astro-ph.GA↗

arXiv2025

CosmoGLINT: Cosmological Generative Model for Line Intensity Mapping with Transformer

Modelling star-forming galaxies is crucial for upcoming observations of large-scale matter and galaxy distributions with galaxy redshift surveys and line intensity mapping (LIM). We introduce CosmoGLINT (Cosmological Generative model for Line INtensity mapping with Transformer), a Transformer-based generative framework designed to create realistic galaxy populations from dark matter (DM)-only simulations. CosmoGLINT auto-regressively generates sequences of galaxy properties -- including star formation rate (SFR), distance to the halo centre, and radial and tangential velocities relative to the halo -- conditioned on halo mass. Trained on the IllustrisTNG hydrodynamic simulation, the model reproduces key statistical properties of the original data, including the voxel intensity distribution and the power spectrum both in real and redshift space. It can efficiently generate a number of different realisations of the designated galaxy populations, enabling the creation of mock LIM/redshift survey catalogues from large halo catalogues produced by fast DM-only simulations. We show that our model trained at multiple redshifts can be applied to DM halo lightcone data to generate a realistic mock galaxy lightcone with incorporating the redshift evolution of the galaxy population. The mock catalogues can be readily used to derive statistical quantities and to develop data analysis pipelines for ongoing and future wide-field surveys.

astro-ph.CO↗

arXiv2025

Scatter in the star formation rate-halo mass relation: secondary bias and its impact on line-intensity mapping

We use the IllustrisTNG cosmological hydrodynamical simulations to study the impact of secondary bias -- specifically, the correlation between star formation rate (SFR) and halo bias at fixed halo mass -- on the line-intensity mapping (LIM) power spectrum. In LIM, the galaxy contributions are flux-weighted, and therefore depend on the luminosity of emission line. We show that the (ensemble-averaged) large-scale two-halo term of the power spectrum depends only on the mean luminosity-halo mass relation if the scatter is uncorrelated with halo bias. However, when luminosity correlates with halo bias at fixed mass, this assumption breaks down. For many emission lines (e.g. H$α$), luminosity is strongly correlated with SFR, making the SFR-weighted power spectrum important to study. In IllustrisTNG, secondary bias increases the two-halo term of the SFR-weighted power spectrum by 5 per cent at $z \sim 1.5$ compared to a model with random scatter. We also find that SFRs of central and satellite galaxies are correlated, enhancing the one-halo term -- which depends on the distribution of SFR inside the halo -- by 10 per cent relative to random pairings. To mitigate secondary bias in the two-halo term, we identify halo concentration (for haloes with mass $\log M_h \lesssim 12$) and satellite mass (for $\log M_h \gtrsim 12$) as effective secondary parameters. These results highlight the need to account for secondary bias when building mock catalogues and interpreting LIM observations.

astro-ph.GA↗

arXiv2025

The First Star-by-star $N$-body/Hydrodynamics Simulation of Our Galaxy Coupling with a Surrogate Model

A major goal of computational astrophysics is to simulate the Milky Way Galaxy with sufficient resolution down to individual stars. However, the scaling fails due to some small-scale, short-timescale phenomena, such as supernova explosions. We have developed a novel integration scheme of $N$-body/hydrodynamics simulations working with machine learning. This approach bypasses the short timesteps caused by supernova explosions using a surrogate model, thereby improving scalability. With this method, we reached 300 billion particles using 148,900 nodes, equivalent to 7,147,200 CPU cores, breaking through the billion-particle barrier currently faced by state-of-the-art simulations. This resolution allows us to perform the first star-by-star galaxy simulation, which resolves individual stars in the Milky Way Galaxy. The performance scales over $10^4$ CPU cores, an upper limit in the current state-of-the-art simulations using both A64FX and X86-64 processors and NVIDIA CUDA GPUs.

astro-ph.GA↗

arXiv2025

ASURA-FDPS-ML: Star-by-star Galaxy Simulations Accelerated by Surrogate Modeling for Supernova Feedback

We introduce new high-resolution galaxy simulations accelerated by a surrogate model that reduces the computation cost by approximately 75 percent. Massive stars with a Zero Age Main Sequence mass of more than about 10 $\mathrm{M_\odot}$ explode as core-collapse supernovae (CCSNe), which play a critical role in galaxy formation. The energy released by CCSNe is essential for regulating star formation and driving feedback processes in the interstellar medium (ISM). However, the short integration timesteps required for SNe feedback have presented significant bottlenecks in astrophysical simulations across various scales. Overcoming this challenge is crucial for enabling star-by-star galaxy simulations, which aim to capture the dynamics of individual stars and the inhomogeneous shell's expansion within the turbulent ISM. To address this, our new framework combines direct numerical simulations and surrogate modeling, including machine learning and Gibbs sampling. The star formation history and the time evolution of outflow rates in the galaxy match those obtained from resolved direct numerical simulations. Our new approach achieves high-resolution fidelity while reducing computational costs, effectively bridging the physical scale gap and enabling multi-scale simulations.

astro-ph.GA↗

arXiv2024

Convolutional Vision Transformer for Cosmology Parameter Inference

Parameter inference is a crucial task in modern cosmology that requires accurate and fast computational methods to handle the high precision and volume of observational datasets. In this study, we explore a hybrid vision transformer, the Convolution vision Transformer (CvT), which combines the benefits of vision transformers (ViTs) and convolutional neural networks (CNNs). We use this approach to infer the $Ω_m$ and $σ_8$ cosmological parameters from simulated dark matter and halo fields. Our experiments indicate that the constraints on $Ω_m$ and $σ_8$ obtained using CvT are better than ViT and CNN, using either dark matter or halo fields. For CvT, pretraining on dark matter fields proves advantageous for improving constraints using halo fields compared to training a model from the beginning. However, ViT and CNN do not show these benefits. The CvT is more efficient than ViT since, despite having more parameters, it requires a training time similar to that of ViT and has similar inference times. The code is available at \url{https://github.com/Yash-10/cvt-cosmo-inference/}.

astro-ph.IM↗

arXiv2024

Insights into the 21 cm field from the vanishing cross-power spectrum at the epoch of reionization

The early stages of the Epoch of Reionization, probed by the 21 cm line, are sensitive to the detailed properties and formation histories of the first galaxies. We use 21cmFAST and a simple, self-consistent galaxy model to examine the redshift evolution of the large-scale cross-power spectrum between the 21 cm field and line-emitting galaxies. A key transition in redshift occurs when the 21 cm field shifts from being positively correlated with the galaxy distribution to being negatively correlated. Importantly, this transition redshift is insensitive to the properties of the galaxy tracers but depends sensitively on the thermal and ionization histories traced through the 21 cm field. Specifically, we show that the transition occurs when both ionization fluctuations dominate over 21 cm spin temperature fluctuations and when the average spin temperature exceeds the temperature of the cosmic microwave background. We illustrate this with three different 21 cm models which have largely the same neutral fraction evolution but different heating histories. We find that the transition redshift has a scale dependence, and that this can help disentangle the relative importance of heating and ionization fluctuations. The best prospects for constraining the transition redshift occur in scenarios with late X-ray heating, where the transition occurs at redshifts as low as $z \sim 6-8$. In our models, this requires high-redshift galaxy surveys with sensitivities of $\sim 10^{-18}~\rm erg/s/cm^2$ for optical lines and $\sim 10^{-19}~\rm erg/s/cm^2$ for FIR lines. Future measurements of the transition redshift can help discriminate between 21 cm models and will benefit from reduced systematics.

astro-ph.CO↗

arXiv2023

Surrogate Modeling for Computationally Expensive Simulations of Supernovae in High-Resolution Galaxy Simulations

Some stars are known to explode at the end of their lives, called supernovae (SNe). The substantial amount of matter and energy that SNe release provides significant feedback to star formation and gas dynamics in a galaxy. SNe release a substantial amount of matter and energy to the interstellar medium, resulting in significant feedback to star formation and gas dynamics in a galaxy. While such feedback has a crucial role in galaxy formation and evolution, in simulations of galaxy formation, it has only been implemented using simple {\it sub-grid models} instead of numerically solving the evolution of gas elements around SNe in detail due to a lack of resolution. We develop a method combining machine learning and Gibbs sampling to predict how a supernova (SN) affects the surrounding gas. The fidelity of our model in the thermal energy and momentum distribution outperforms the low-resolution SN simulations. Our method can replace the SN sub-grid models and help properly simulate un-resolved SN feedback in galaxy formation simulations. We find that employing our new approach reduces the necessary computational cost to $\sim$ 1 percent compared to directly resolving SN feedback.

astro-ph.GA↗

arXiv2023

3D-Spatiotemporal Forecasting the Expansion of Supernova Shells Using Deep Learning toward High-Resolution Galaxy Simulations

Supernova (SN) plays an important role in galaxy formation and evolution. In high-resolution galaxy simulations using massively parallel computing, short integration timesteps for SNe are serious bottlenecks. This is an urgent issue that needs to be resolved for future higher-resolution galaxy simulations. One possible solution would be to use the Hamiltonian splitting method, in which regions requiring short timesteps are integrated separately from the entire system. To apply this method to the particles affected by SNe in a smoothed-particle hydrodynamics simulation, we need to detect the shape of the shell on and within which such SN-affected particles reside during the subsequent global step in advance. In this paper, we develop a deep learning model, 3D-MIM, to predict a shell expansion after a SN explosion. Trained on turbulent cloud simulations with particle mass $m_{\rm gas}=1$M$_\odot$, the model accurately reproduces the anisotropic shell shape, where densities decrease by over 10 per cent by the explosion. We also demonstrate that the model properly predicts the shell radius in the uniform medium beyond the training dataset of inhomogeneous turbulent clouds. We conclude that our model enables the forecast of the shell and its interior where SN-affected particles will be present.

astro-ph.GA↗

arXiv2023

The 300 pc resolution imaging of a z = 8.31 galaxy: Turbulent ionized gas and potential stellar feedback 600 million years after the Big Bang

We present the results of 300 pc resolution ALMA imaging of the [OIII] 88 $μ$m line and dust continuum emission from a $z = 8.312$ Lyman break galaxy MACS0416_Y1. The velocity-integrated [OIII] emission has three peaks which are likely associated with three young stellar clumps of MACS0416_Y1, while the channel map shows a complicated velocity structure with little indication of a global velocity gradient unlike what was found in [CII] 158 $μ$m at a larger scale, suggesting random bulk motion of ionized gas clouds inside the galaxy. In contrast, dust emission appears as two individual clumps apparently separating or bridging the [OIII]/stellar clumps. The cross correlation coefficient between dust and ultraviolet-related emission (i.e., [OIII] and ultraviolet continuum) is unity on a galactic scale, while it drops at < 1 kpc, suggesting well mixed geometry of multi-phase interstellar media on sub-kpc scales. If the cutoff scale characterizes different stages of star formation, the cutoff scale can be explained by gravitational instability of turbulent gas. We also report on a kpc-scale off-center cavity embedded in the dust continuum image. This could be a superbubble producing galactic-scale outflows, since the energy injection from the 4 Myr starburst suggested by a spectral energy distribution analysis is large enough to push the surrounding media creating a kpc-scale cavity.

astro-ph.GA↗

arXiv2023

Machine Learning for Observational Cosmology

An array of large observational programs using ground-based and space-borne telescopes is planned in the next decade. The forthcoming wide-field sky surveys are expected to deliver a sheer volume of data exceeding an exabyte. Processing the large amount of multiplex astronomical data is technically challenging, and fully automated technologies based on machine learning and artificial intelligence are urgently needed. Maximizing scientific returns from the big data requires community-wide efforts. We summarize recent progress in machine learning applications in observational cosmology. We also address crucial issues in high-performance computing that are needed for the data processing and statistical analysis.

astro-ph.IM↗

arXiv2022

Possible Systematic Rotation in the Mature Stellar Population of a $z=9.1$ Galaxy

We present new observations with the Atacama Large Millimeter/submillimeter Array for a gravitationally-lensed galaxy at $z=9.1$, MACS1149-JD1. [O III] 88-$μ$m emission is detected at 10$σ$ with a spatial resolution of $\sim0.3$ kpc in the source plane, enabling the most distant morpho-kinematic study of a galaxy. The [O III] emission is distributed smoothly without any resolved clumps and shows a clear velocity gradient with $ΔV_{\rm obs}/2σ_{\rm tot}=0.84\pm0.23$, where $ΔV_{\rm obs}$ is the observed maximum velocity difference and $σ_{\rm tot}$ is the velocity dispersion measured in the spatially-integrated line profile, suggesting a rotating system. Assuming a geometrically thin self-gravitating rotation disk model, we obtain $V_{\rm rot}/σ_V=0.67_{-0.26}^{+0.73}$, where $V_{\rm rot}$ and $σ_V$ are the rotation velocity and velocity dispersion, respectively, still consistent with rotation. The resulting disk mass of $0.65_{-0.40}^{+1.37}\times10^{9}$ M$_\odot$ is consistent with being associated with the stellar mass identified with a 300 Myr-old stellar population independently indicated by a Balmer break in the spectral energy distribution. We conclude that the most of the dynamical mass is associated with the previously-identified mature stellar population that formed at $z\sim15$.

astro-ph.GA↗

arXiv2021

Deep learning reconstruction of three-dimensional galaxy distributions with intensity mapping observations

Line intensity mapping is emerging as a novel method that can measure the collective intensity fluctuations of atomic/molecular line emission from distant galaxies. Several observational programs with various wavelengths are ongoing and planned, but there remains a critical problem of line confusion; emission lines originating from galaxies at different redshifts are confused at the same observed wavelength. We devise a generative adversarial network that extracts designated emission line signals from noisy three-dimensional data. Our novel network architecture allows two input data, in which the same underlying large-scale structure is traced by two emission lines of H$\rm α$ and [OIII], so that the network learns the relative contributions at each wavelength and is trained to decompose the respective signals. After being trained with a large number of realistic mock catalogs, the network is able to reconstruct the three-dimensional distribution of emission-line galaxies at $z = 1.3-2.4$. Bright galaxies are identified with a precision of 84%, and the cross-correlation coefficients between the true and reconstructed intensity maps are as high as 0.8. Our deep-learning method can be readily applied to data from planned space-borne and ground-based experiments.

astro-ph.GA↗

arXiv2021

Noise reduction for weak lensing mass mapping: An application of generative adversarial networks to Subaru Hyper Suprime-Cam first-year data

We propose a deep-learning approach based on generative adversarial networks (GANs) to reduce noise in weak lensing mass maps under realistic conditions. We apply image-to-image translation using conditional GANs to the mass map obtained from the first-year data of Subaru Hyper Suprime-Cam (HSC) survey. We train the conditional GANs by using 25000 mock HSC catalogues that directly incorporate a variety of observational effects. We study the non-Gaussian information in denoised maps using one-point probability distribution functions (PDFs) and also perform matching analysis for positive peaks and massive clusters. An ensemble learning technique with our GANs is successfully applied to reproduce the PDFs of the lensing convergence. About $60\%$ of the peaks in the denoised maps with height greater than $5σ$ have counterparts of massive clusters within a separation of 6 arcmin. We show that PDFs in the denoised maps are not compromised by details of multiplicative biases and photometric redshift distributions, nor by shape measurement errors, and that the PDFs show stronger cosmological dependence compared to the noisy counterpart. We apply our denoising method to a part of the first-year HSC data to show that the observed mass distribution is statistically consistent with the prediction from the standard $Λ$CDM model.

astro-ph.CO↗