Subscribe to the gold package and get unlimited access to Shamra Academy
Register a new userShear measurement bias II: a fast machine learning calibration method
120
0
0.0
(
0
)
Ask ChatGPT about the research
No Arabic abstract
We present a new shear calibration method based on machine learning. The method estimates the individual shear responses of the objects from the combination of several measured properties on the images using supervised learning. The supervised learning uses the true individual shear responses obtained from copies of the image simulations with different shear values. On simulated GREAT3data, we obtain a residual bias after the calibration compatible with 0 and beyond Euclid requirements for a signal-to-noise ratio > 20 within ~15 CPU hours of training using only ~10^5 objects. This efficient machine-learning approach can use a smaller data set because the method avoids the contribution from shape noise. The low dimensionality of the input data also leads to simple neural network architectures. We compare it to the recently described method Metacalibration, which shows similar performances. The different methods and systematics suggest that the two methods are very good complementary methods. Our method can therefore be applied without much effort to any survey such as Euclid or the Vera C. Rubin Observatory, with fewer than a million images to simulate to learn the calibration function.
rate research
Read More
Sample selection is a necessary preparation for weak lensing measurement. It is well-known that selection itself may introduce bias in the measured shear signal. Using image simulation and the Fourier_Quad shear measurement pipeline, we quantify the selection bias in various commonly used selection function (signal-to-noise-ratio, magnitude, etc.). We proposed a new selection function defined in the power spectrum of the galaxy image. This new selection function has low selection bias, and it is particularly convenient for shear measurement pipelines based on Fourier transformation.
Accurate shape measurements are essential to infer cosmological parameters from large area weak gravitational lensing studies. The compact diffraction-limited point-spread function (PSF) in space-based observations is greatly beneficial, but its chromaticity for a broad band observation can lead to new subtle effects that could hitherto be ignored: the PSF of a galaxy is no longer uniquely defined and spatial variations in the colours of galaxies result in biases in the inferred lensing signal. Taking Euclid as a reference, we show that this colourgradient bias (CG bias) can be quantified with high accuracy using available multi-colour Hubble Space Telescope (HST) data. In particular we study how noise in the HST observations might impact such measurements and find this to be negligible. We determine the CG bias using HST observations in the F606W and F814W filters and observe a correlation with the colour, in line with expectations, whereas the dependence with redshift is weak. The biases for individual galaxies are generally well below 1%, which may be reduced further using morphological information from the Euclid data. Our results demonstrate that CG bias should not be ignored, but it is possible to determine its amplitude with sufficient precision, so that it will not significantly bias the weak lensing measurements using Euclid data.
With the advent of large-scale weak lensing surveys there is a need to understand how realistic, scale-dependent systematics bias cosmic shear and dark energy measurements, and how they can be removed. Here we describe how spatial variations in the amplitude and orientation of realistic image distortions convolve with the measured shear field, mixing the even-parity convergence and odd-parity modes, and bias the shear power spectrum. Many of these biases can be removed by calibration to external data, the survey itself, or by modelling in simulations. The uncertainty in the calibration must be marginalised over and we calculate how this propagates into parameter estimation, degrading the dark energy Figure-of-Merit. We find that noise-like biases affect dark energy measurements the most, while spikes in the bias power have the least impact, reflecting their correlation with the effect of cosmological parameters. We argue that in order to remove systematic biases in cosmic shear surveys and maintain statistical power effort should be put into improving the accuracy of the bias calibration rather than minimising the size of the bias. In general, this appears to be a weaker condition for bias removal. We also investigate how to minimise the size of the calibration set for a fixed reduction in the Figure-of-Merit. These results can be used to model the effect of biases and calibration on a cosmic shear survey accurately, assess their impact on the measurement of modified gravity and dark energy models, and to optimise surveys and calibration requirements.
Forthcoming large-scale surveys will soon attempt to measure cosmic shear to an unprecedented level of accuracy, requiring a similarly high level of accuracy in the shear measurements of galaxies. Factors such as pixelisation, imperfect point-spread function (PSF) correction, and pixel noise can all directly or indirectly lead to biases in shear measurements, and so it can be necessary for shear measurement methods to be calibrated against internal, external, or simulated data to minimize bias. It is thus important to understand the nature of this calibration. In this paper, we show that a typical calibration procedure will on average leave no residual additive bias, but will leave a residual multiplicative bias. Additionally, the errors on the post-calibration bias parameters will be changed, and on average increased, from the errors on the pre-calibration measurements of these parameters, but that this is generally worth the benefit in decreasing the expected value of the multiplicative bias. We find that in most typical cases, it is worthwhile to apply a first-order bias correction, while a higher-order bias correction is only worthwhile for methods with intrinsically high multiplicative bias ($>10$ per cent) or when the simulation size is very small ($<10^6$ simulated galaxies).
We present results from a set of simulations designed to constrain the weak lensing shear calibration for the Hyper Suprime-Cam (HSC) survey. These simulations include HSC observing conditions and galaxy images from the Hubble Space Telescope (HST), with fully realistic galaxy morphologies and the impact of nearby galaxies included. We find that the inclusion of nearby galaxies in the images is critical to reproducing the observed distributions of galaxy sizes and magnitudes, due to the non-negligible fraction of unrecognized blends in ground-based data, even with the excellent typical seeing of the HSC survey (0.58 in the $i$-band). Using these simulations, we detect and remove the impact of selection biases due to the correlation of weights and the quantities used to define the sample (S/N and apparent size) with the lensing shear. We quantify and remove galaxy property-dependent multiplicative and additive shear biases that are intrinsic to our shear estimation method, including a $sim 10$ per cent-level multiplicative bias due to the impact of nearby galaxies and unrecognized blends. Finally, we check the sensitivity of our shear calibration estimates to other cuts made on the simulated samples, and find that the changes in shear calibration are well within the requirements for HSC weak lensing analysis. Overall, the simulations suggest that the weak lensing multiplicative biases in the first-year HSC shear catalog are controlled at the 1 per cent level.
Log in to be able to interact and post comments
comments
Fetching comments
Sign in to be able to follow your search criteria
