Dealing with missing data in the MICROSCOPE space mission: An adaptation of inpainting to handle colored-noise data

Authors: S. Pires, B. Joël, Q. Baghi, P. Touboul, G. Metris
Journal: Physical Review D
Year: 2016
Download: ADS | arXiv


Abstract

The MICROSCOPE space mission, launched on April 25, 2016, aims to test the weak equivalence principle (WEP) with a 10-15 precision. Reaching this performance requires an accurate and robust data analysis method, especially since the possible WEP violation signal will be dominated by a strongly colored noise. An important complication is brought by the fact that some values will be missing—therefore, the measured time series will not be strictly regularly sampled. Those missing values induce a spectral leakage that significantly increases the noise in Fourier space, where the WEP violation signal is looked for, thereby complicating scientific returns. Recently, we developed an inpainting algorithm to correct the MICROSCOPE data for missing values. This code has been integrated in the official MICROSCOPE data processing and analysis pipeline because it enables us to significantly measure an equivalence principle violation (EPV) signal in a model-independent way, in the inertial satellite configuration. In this work, we present several improvements to the method that may allow us now to reach the MICROSCOPE requirements for both inertial and spin satellite configurations. The main improvement has been obtained using a prior on the power spectrum of the colored noise that can be directly derived from the incomplete data. We show that after reconstructing missing values with this new algorithm, a least-squares fit may allow us to significantly measure an EPV signal with a 0.96 ×10-15 precision in the inertial mode and 1.20 ×10-15 precision in the spin mode. Although, the inpainting method presented in this paper has been optimized to the MICROSCOPE data, it remains sufficiently general to be used in the general context of missing data in time series dominated by an unknown colored noise. The improved inpainting software, called inpainting for colored-noise dominated signals, is freely available at http://www.cosmostat.org/software/icon.

High Resolution Weak Lensing Mass-Mapping Combining Shear and Flexion

Authors: F. Lanusse, J.-L. Starck, A. Leonard, S. Pires
Journal: A&A
Year: 2016
Download: ADS | arXiv


Abstract

Aims: We propose a new mass mapping algorithm, specifically designed to recover small-scale information from a combination of gravitational shear and flexion. Including flexion allows us to supplement the shear on small scales in order to increase the sensitivity to substructures and the overall resolution of the convergence map without relying on strong lensing constraints.
Methods: To preserve all available small scale information, we avoid any binning of the irregularly sampled input shear and flexion fields and treat the mass mapping problem as a general ill-posed inverse problem, which is regularised using a robust multi-scale wavelet sparsity prior. The resulting algorithm incorporates redshift, reduced shear, and reduced flexion measurements for individual galaxies and is made highly efficient by the use of fast Fourier estimators.
Results: We tested our reconstruction method on a set of realistic weak lensing simulations corresponding to typical HST/ACS cluster observations and demonstrate our ability to recover substructures with the inclusion of flexion, which are otherwise lost if only shear information is used. In particular, we can detect substructures on the 15'' scale well outside of the critical region of the clusters. In addition, flexion also helps to constrain the shape of the central regions of the main dark matter halos.

A new model to predict weak-lensing peak counts III. Filtering technique comparisons

Authors: C. Lin, M. Kilbinger, S. Pires
Journal: A&A
Year: 2016
Download: ADS | arXiv


Abstract

This is the third in a series of papers that develop a new and flexible model to predict weak-lensing (WL) peak counts, which have been shown to be a very valuable non-Gaussian probe of cosmology. In this paper, we compare the cosmological information extracted from WL peak counts using different filtering techniques of the galaxy shear data, including linear filtering with a Gaussian and two compensated filters (the starlet wavelet and the aperture mass), and the nonlinear filtering method MRLens. We present improvements to our model that account for realistic survey conditions, which are masks, shear-to-convergence transformations, and non-constant noise. We create simulated peak counts from our stochastic model, from which we obtain constraints on the matter density Ωm, the power spectrum normalisation σ8, and the dark-energy parameter w0. We use two methods for parameter inference, a copula likelihood, and approximate Bayesian computation (ABC). We measure the contour width in the Ωm-σ8 degeneracy direction and the figure of merit to compare parameter constraints from different filtering techniques. We find that starlet filtering outperforms the Gaussian kernel, and that including peak counts from different smoothing scales helps to lift parameter degeneracies. Peak counts from different smoothing scales with a compensated filter show very little cross-correlation, and adding information from different scales can therefore strongly enhance the available information. Measuring peak counts separately from different scales yields tighter constraints than using a combined peak histogram from a single map that includes multiscale information. Our results suggest that a compensated filter function with counts included separately from different smoothing scales yields the tightest constraints on cosmological parameters from WL peaks.

Sparsely sampling the sky: Regular vs. random sampling

 

Authors: P. Paykari, S. Pires, J.-L. Starck, A.H. Jaffe
Journal: Astronomy & Astrophysics
Year: 2015
Download: ADS | arXiv


Abstract

Weak gravitational lensing provides a unique way of mapping directly the dark matter in the Universe. The majority of lensing analyses use the two-point statistics of the cosmic shear field to constrain the cosmological model, a method that is affected by degeneracies, such as that between σ8 and Ωm which are respectively the rms of the mass fluctuations on a scale of 8 Mpc/h and the matter density parameter, both at z = 0. However, the two-point statistics only measure the Gaussian properties of the field, and the weak lensing field is non-Gaussian. It has been shown that the estimation of non-Gaussian statistics for weak lensing data can improve the constraints on cosmological parameters. In this paper, we systematically compare a wide range of non-Gaussian estimators to determine which one provides tighter constraints on the cosmological parameters. These statistical methods include skewness, kurtosis, and the higher criticism test, in several sparse representations such as wavelet and curvelet; as well as the bispectrum, peak counting, and a newly introduced statistic called wavelet peak counting (WPC). Comparisons based on sparse representations indicate that the wavelet transform is the most sensitive to non-Gaussian cosmological structures. It also appears that the most helpful statistic for non-Gaussian characterization in weak lensing mass maps is the WPC. Finally, we show that the σ8 - Ωmdegeneracy could be even better broken if the WPC estimation is performed on weak lensing mass maps filtered by the wavelet method, MRLens.

Dealing with missing data: An inpainting application to the MICROSCOPE space mission

Authors: B. Joël, S. Pires, Q. Baghi, P. Touboul, G. Metris
Journal: Physical Review D
Year: 2015
Download: ADS | arXiv


Abstract

Missing data are a common problem in experimental and observational physics. They can be caused by various sources, either an instrument's saturation, or a contamination from an external event, or a data loss. In particular, they can have a disastrous effect when one is seeking to characterize a colored-noise-dominated signal in Fourier space, since they create a spectral leakage that can artificially increase the noise. It is therefore important to either take them into account or to correct for them prior to e.g. a Least-Square fit of the signal to be characterized. In this paper, we present an application of the {\it inpainting} algorithm to mock MICROSCOPE data; {\it inpainting} is based on a sparsity assumption, and has already been used in various astrophysical contexts; MICROSCOPE is a French Space Agency mission, whose launch is expected in 2016, that aims to test the Weak Equivalence Principle down to the 1015 level. We then explore the {\it inpainting} dependence on the number of gaps and the total fraction of missing values. We show that, in a worst-case scenario, after reconstructing missing values with {\it inpainting}, a Least-Square fit may allow us to significantly measure a 1.1×1015 Equivalence Principle violation signal, which is sufficiently close to the MICROSCOPE requirements to implement {\it inpainting} in the official MICROSCOPE data processing and analysis pipeline. Together with the previously published KARMA method, {\it inpainting} will then allow us to independently characterize and cross-check an Equivalence Principle violation signal detection down to the 1015 level.

A PCA-based automated finder for galaxy-scale strong lenses

 

Authors: R. Joseph, F. Courbin, R. B. Metcalf, ...., S.Pires, et al.
Journal: A&A
Year: 2014
Download: ADS | arXiv


Abstract

We present an algorithm using principal component analysis (PCA) to subtract galaxies from imaging data and also two algorithms to find strong, galaxy-scale gravitational lenses in the resulting residual image. The combined method is optimised to find full or partial Einstein rings. Starting from a pre-selection of potential massive galaxies, we first perform a PCA to build a set of basis vectors. The galaxy images are reconstructed using the PCA basis and subtracted from the data. We then filter the residual image with two different methods. The first uses a curvelet (curved wavelets) filter of the residual images to enhance any curved/ring feature. The resulting image is transformed in polar coordinates, centred on the lens galaxy. In these coordinates, a ring is turned into a line, allowing us to detect very faint rings by taking advantage of the integrated signal-to-noise in the ring (a line in polar coordinates). The second way of analysing the PCA-subtracted images identifies structures in the residual images and assesses whether they are lensed images according to their orientation, multiplicity, and elongation. We applied the two methods to a sample of simulated Einstein rings as they would be observed with the ESA Euclid satellite in the VIS band. The polar coordinate transform allowed us to reach a completeness of 90% for a purity of 86%, as soon as the signal-to-noise integrated in the ring was higher than 30 and almost independent of the size of the Einstein ring. Finally, we show with real data that our PCA-based galaxy subtraction scheme performs better than traditional subtraction based on model fitting to the data. Our algorithm can be developed and improved further using machine learning and dictionary learning methods, which would extend the capabilities of the method to more complex and diverse galaxy shapes.

Gap interpolation by inpainting methods : Application to Ground and Space-based Asteroseismic data

 

Authors: S. Pires, S. Mathur, R. A. Garcia, J. Ballot, D. Stello, K. Sato
Journal: Astronomy & Astrophysics
Year: 2014
Download: ADS | arXiv


Abstract

In asteroseismology, the observed time series often suffers from incomplete time coverage due to gaps. The presence of periodic gaps may generate spurious peaks in the power spectrum that limit the analysis of the data. Various methods have been developed to deal with gaps in time series data. However, it is still important to improve these methods to be able to extract all the possible information contained in the data. In this paper, we propose a new approach to handle the problem, the so-called inpainting method. This technique, based on a sparsity prior, enables to judiciously fill-in the gaps in the data, preserving the asteroseismic signal, as far as possible. The impact of the observational window function is reduced and the interpretation of the power spectrum is simplified. This method is applied both on ground and space-based data. It appears that the inpainting technique improves the oscillation modes detection and estimation. Additionally, it can be used to study very long time series of many stars because its computation is very fast. For a time series of 50 days of CoRoT-like data, it allows a speed-up factor of 1000, if compared to methods of the same accuracy.


Summary

The paper "Gap interpolation by inpainting methods : Application to Ground and Space-based Asteroseismic data" has been accepted for publication in A&A. This paper is describing the software K-inpainting.

The software K-inpainting has been developed to handle the problem of missing data in the asteroseismic signal of Kepler. This technique, based on a sparsity prior, enables to judiciously fill-in the gaps in the data, preserving the asteroseismic signal, as far as possible.

More recently, it has been implemented in the CoRoT pipeline to process the data from missing data.

The K-inapinting software is available here

3346_3

Power Density Spectrum for a duty cycle of 83% computed using an FFT on the inpainted time series.

Impact on asteroseismic analyses of regular gaps in Kepler data

 

Authors: R.A. Garcıa, S. Mathur, S. Pires, et al.
Journal: Astronomy & Astrophysics
Year: 2014
Download: ADS | arXiv


Abstract

The NASA Kepler mission has observed more than 190,000 stars in the constellations of Cygnus and Lyra. Around 4 years of almost continuous ultra high-precision photometry have been obtained reaching a duty cycle higher than 90% for many of these stars. However, almost regular gaps due to nominal operations are present in the light curves at different time scales. In this paper we want to highlight the impact of those regular gaps in asteroseismic analyses and we try to find a method that minimizes their effect in the frequency domain. To do so, we isolate the two main time scales of quasi regular gaps in the data. We then interpolate the gaps and we compare the power density spectra of four different stars: two red giants at different stages of their evolution, a young F-type star, and a classical pulsator in the instability strip. The spectra obtained after filling the gaps in the selected solar-like stars show a net reduction in the overall background level, as well as a change in the background parameters. The inferred convective properties could change as much as 200% in the selected example, introducing a bias in the p-mode frequency of maximum power. When global asteroseismic scaling relations are used, this bias can lead up to a variation in the surface gravity of 0.05 dex. Finally, the oscillation spectrum in the classical pulsator is cleaner compared to the original one.

Weak Lensing Galaxy Cluster Field Reconstruction

 

Authors: E. Jullo, S.Pires, M. Jauzac, J.-P. Kneib
Journal: MNRAS
Year: 2014
Download: ADS | arXiv


Abstract

In this paper, we compare three methods to reconstruct galaxy cluster density fields with weak lensing data. The first method called FLens integrates an inpainting concept to invert the shear field with possible gaps, and a multi-scale entropy denoising procedure to remove the noise contained in the final reconstruction, that arises mostly from the random intrinsic shape of the galaxies. The second and third methods are based on a model of the density field made of a multi-scale grid of radial basis functions. In one case, the model parameters are computed with a linear inversion involving a singular value decomposition. In the other case, the model parameters are estimated using a Bayesian MCMC optimization implemented in the lensing software Lenstool. Methods are compared on simulated data with varying galaxy density fields. We pay particular attention to the errors estimated with resampling. We find the multi-scale grid model optimized with MCMC to provide the best results, but at high computational cost, especially when considering resampling. The SVD method is much faster but yields noisy maps, although this can be mitigated with resampling. The FLens method is a good compromise with fast computation, high signal to noise reconstruction, but lower resolution maps. All three methods are applied to the MACS J0717+3745 galaxy cluster field, and reveal the filamentary structure discovered in Jauzac et al. 2012. We conclude that sensitive priors can help to get high signal to noise, and unbiased reconstructions.

Defining a weak lensing experiment in space

 

Authors: M. Cropper, H. Hoekstra, T. Kitching, ..., S. Pires et al.
Journal: MNRAS
Year: 2013
Download: ADS | arXiv


Abstract

This paper describes the definition of a typical next-generation space-based weak gravitational lensing experiment. We first adopt a set of top-level science requirements from the literature, based on the scale and depth of the galaxy sample, and the avoidance of systematic effects in the measurements which would bias the derived shear values. We then identify and categorise the contributing factors to the systematic effects, combining them with the correct weighting, in such a way as to fit within the top-level requirements. We present techniques which permit the performance to be evaluated and explore the limits at which the contributing factors can be managed. Besides the modelling biases resulting from the use of weighted moments, the main contributing factors are the reconstruction of the instrument point spread function (PSF), which is derived from the stellar images on the image, and the correction of the charge transfer inefficiency (CTI) in the CCD detectors caused by radiation damage.

If instrumentation is stable and well calibrated, we find extant shear measurement software from Gravitational Lensing Accuracy Testing 2010 (GREAT10) already meet requirements on galaxies detected at signal-to-noise ratio = 40. Averaging over a population of galaxies with a realistic distribution of sizes, it also meets requirements for a 2D cosmic shear analysis from space. If used on fainter galaxies or for 3D cosmic shear tomography, existing algorithms would need calibration on simulations to avoid introducing bias at a level similar to the statistical error. Requirements on hardware and calibration data are discussed in more detail in a companion paper. Our analysis is intentionally general, but is specifically being used to drive the hardware and ground segment performance budget for the design of the European Space Agency's recently selected Euclid mission.