The Universe is worth $64^3$ pixels: Convolution Neural Network and Vision Transformers for Cosmology
Authors:
Se Yeon Hwang,
Cristiano G. Sabiu,
Inkyu Park,
Sungwook E. Hong
Abstract:
We present a novel approach for estimating cosmological parameters, $Ω_m$, $σ_8$, $w_0$, and one derived parameter, $S_8$, from 3D lightcone data of dark matter halos in redshift space covering a sky area of $40^\circ \times 40^\circ$ and redshift range of $0.3 < z < 0.8$, binned to $64^3$ voxels. Using two deep learning algorithms, Convolutional Neural Network (CNN) and Vision Transformer (ViT),…
▽ More
We present a novel approach for estimating cosmological parameters, $Ω_m$, $σ_8$, $w_0$, and one derived parameter, $S_8$, from 3D lightcone data of dark matter halos in redshift space covering a sky area of $40^\circ \times 40^\circ$ and redshift range of $0.3 < z < 0.8$, binned to $64^3$ voxels. Using two deep learning algorithms, Convolutional Neural Network (CNN) and Vision Transformer (ViT), we compare their performance with the standard two-point correlation (2pcf) function. Our results indicate that CNN yields the best performance, while ViT also demonstrates significant potential in predicting cosmological parameters. By combining the outcomes of Vision Transformer, Convolution Neural Network, and 2pcf, we achieved a substantial reduction in error compared to the 2pcf alone. To better understand the inner workings of the machine learning algorithms, we employed the Grad-CAM method to investigate the sources of essential information in activation maps of the CNN and ViT. Our findings suggest that the algorithms focus on different parts of the density field and redshift depending on which parameter they are predicting. This proof-of-concept work paves the way for incorporating deep learning methods to estimate cosmological parameters from large-scale structures, potentially leading to tighter constraints and improved understanding of the Universe.
△ Less
Submitted 2 November, 2023; v1 submitted 17 April, 2023;
originally announced April 2023.
Studying High-Mass Microquasars with HAWC
Authors:
Chang Dong Rho,
Ke Fang,
Se Yeon Hwang,
Youngwan Son
Abstract:
High-mass microquasars (HMMQs) are powerful particle accelerators, but their mechanism of the high-energy emission is poorly understood. To date, only a handful of these particle engines have ever been observed to emit gamma-ray photons and are thus potential TeV gamma-ray emitters. In this work, we study four HMMQs, namely, LS 5039, Cyg X-1, Cyg X-3, and SS 433 using the data from the High Altitu…
▽ More
High-mass microquasars (HMMQs) are powerful particle accelerators, but their mechanism of the high-energy emission is poorly understood. To date, only a handful of these particle engines have ever been observed to emit gamma-ray photons and are thus potential TeV gamma-ray emitters. In this work, we study four HMMQs, namely, LS 5039, Cyg X-1, Cyg X-3, and SS 433 using the data from the High Altitude Water Cherenkov (HAWC) observatory. We perform time dependent analyses on each HMMQ to look for any periodic variations in their flux. We produce light curves using the HAWC daily maps from which we generate Lomb-Scargle periodograms. By analysing the significance of the periodogram peaks, we assess whether or not HAWC is sensitive to orbitally modulating TeV gamma-ray flux in the four HMMQs.
△ Less
Submitted 1 August, 2021;
originally announced August 2021.
