arXiv subjects

Joel Akeret

Publications and source records attributed to Joel Akeret.

7 recordsLinked to original sources

arXiv2017

Radio frequency interference mitigation using deep convolutional neural networks

We propose a novel approach for mitigating radio frequency interference (RFI) signals in radio data using the latest advances in deep learning. We employ a special type of Convolutional Neural Network, the U-Net, that enables the classification of clean signal and RFI signatures in 2D time-ordered data acquired from a radio telescope. We train and assess the performance of this network using the HIDE & SEEK radio data simulation and processing packages, as well as early Science Verification data acquired with the 7m single-dish telescope at the Bleien Observatory. We find that our U-Net implementation is showing competitive accuracy to classical RFI mitigation algorithms such as SEEK's SumThreshold implementation. We publish our U-Net software package on GitHub under GPLv3 license.

astro-ph.IM↗

arXiv2017

HIDE & SEEK: End-to-End Packages to Simulate and Process Radio Survey Data

As several large single-dish radio surveys begin operation within the coming decade, a wealth of radio data will become available and provide a new window to the Universe. In order to fully exploit the potential of these data sets, it is important to understand the systematic effects associated with the instrument and the analysis pipeline. A common approach to tackle this is to forward-model the entire system - from the hardware to the analysis of the data products. For this purpose, we introduce two newly developed, open-source Python packages: the HI Data Emulator (HIDE) and the Signal Extraction and Emission Kartographer (SEEK) for simulating and processing single-dish radio survey data. HIDE forward-models the process of collecting astronomical radio signals in a single-dish radio telescope instrument and outputs pixel-level time-ordered-data. SEEK processes the time-ordered-data, removes artifacts from Radio Frequency Interference (RFI), automatically applies flux calibration, and aims to recover the astronomical radio signal. The two packages can be used separately or together depending on the application. Their modular and flexible nature allows easy adaptation to other instruments and data sets. We describe the basic architecture of the two packages and examine in detail the noise and RFI modeling in HIDE, as well as the implementation of gain calibration and RFI mitigation in SEEK. We then apply HIDE & SEEK to forward-model a Galactic survey in the frequency range 990 - 1260 MHz based on data taken at the Bleien Observatory. For this survey, we expect to cover 70% of the full sky and achieve a median signal-to-noise ratio of approximately 5 - 6 in the cleanest channels including systematic uncertainties. However, we also point out the potential challenges of high RFI contamination and baseline removal when examining the early data from the Bleien Observatory.

astro-ph.IM↗

arXiv2016

An Integrated System at the Bleien Observatory for Mapping the Galaxy

We describe the design and performance of the hardware system at the Bleien Observatory. The system is designed to deliver a map of the Galaxy for studying the foreground contamination of low-redshift (z=0.13--0.43) H$_{\rm I}$ intensity mapping experiments as well as other astronomical Galactic studies. This hardware system is composed of a 7m parabolic dish, a dual-polarization corrugated horn feed, a pseudo correlation receiver, a Fast Fourier Transform spectrometer, and an integrated control system that controls and monitors the progress of the data collection. The main innovative designs in the hardware are (1) the pseudo correlation receiver and the cold reference source within (2) the high dynamic range, high frequency resolution spectrometer and (3) the phase-switch implementation of the system. This is the first time these technologies are used together for a L-band radio telescope to achieve an electronically stable system, which is an essential first step for wide-field cosmological measurements. This work demonstrates the prospects and challenges for future H$_{\rm I}$ intensity mapping experiments.

astro-ph.IM↗

arXiv2016

Simulating the Large-Scale Structure of HI Intensity Maps

Intensity mapping of neutral hydrogen (HI) is a promising observational probe of cosmology and large-scale structure. We present wide field simulations of HI intensity maps based on N-body simulations of a $2.6\, {\rm Gpc / h}$ box with $2048^3$ particles (particle mass $1.6 \times 10^{11}\, {\rm M_\odot / h}$). Using a conditional mass function to populate the simulated dark matter density field with halos below the mass resolution of the simulation ($10^{8}\, {\rm M_\odot / h} < M_{\rm halo} < 10^{13}\, {\rm M_\odot / h}$), we assign HI to those halos according to a phenomenological halo to HI mass relation. The simulations span a redshift range of 0.35 < z < 0.9 in redshift bins of width $Δz \approx 0.05$ and cover a quarter of the sky at an angular resolution of about 7'. We use the simulated intensity maps to study the impact of non-linear effects and redshift space distortions on the angular clustering of HI. Focusing on the autocorrelations of the maps, we apply and compare several estimators for the angular power spectrum and its covariance. We verify that these estimators agree with analytic predictions on large scales and study the validity of approximations based on Gaussian random fields, particularly in the context of the covariance. We discuss how our results and the simulated maps can be useful for planning and interpreting future HI intensity mapping surveys.

astro-ph.CO↗

arXiv2015

Approximate Bayesian Computation for Forward Modeling in Cosmology

Bayesian inference is often used in cosmology and astrophysics to derive constraints on model parameters from observations. This approach relies on the ability to compute the likelihood of the data given a choice of model parameters. In many practical situations, the likelihood function may however be unavailable or intractable due to non-gaussian errors, non-linear measurements processes, or complex data formats such as catalogs and maps. In these cases, the simulation of mock data sets can often be made through forward modeling. We discuss how Approximate Bayesian Computation (ABC) can be used in these cases to derive an approximation to the posterior constraints using simulated data sets. This technique relies on the sampling of the parameter set, a distance metric to quantify the difference between the observation and the simulations and summary statistics to compress the information in the data. We first review the principles of ABC and discuss its implementation using a Population Monte-Carlo (PMC) algorithm and the Mahalanobis distance metric. We test the performance of the implementation using a Gaussian toy model. We then apply the ABC technique to the practical case of the calibration of image simulations for wide field cosmological surveys. We find that the ABC analysis is able to provide reliable parameter constraints for this problem and is therefore a promising technique for other applications in cosmology and astrophysics. Our implementation of the ABC PMC method is made available via a public code release.

astro-ph.CO↗

arXiv2014

HOPE: A Python Just-In-Time compiler for astrophysical computations

The Python programming language is becoming increasingly popular for scientific applications due to its simplicity, versatility, and the broad range of its libraries. A drawback of this dynamic language, however, is its low runtime performance which limits its applicability for large simulations and for the analysis of large data sets, as is common in astrophysics and cosmology. While various frameworks have been developed to address this limitation, most focus on covering the complete language set, and either force the user to alter the code or are not able to reach the full speed of an optimised native compiled language. In order to combine the ease of Python and the speed of C++, we developed HOPE, a specialised Python just-in-time (JIT) compiler designed for numerical astrophysical applications. HOPE focuses on a subset of the language and is able to translate Python code into C++ while performing numerical optimisation on mathematical expressions at runtime. To enable the JIT compilation, the user only needs to add a decorator to the function definition. We assess the performance of HOPE by performing a series of benchmarks and compare its execution speed with that of plain Python, C++ and the other existing frameworks. We find that HOPE improves the performance compared to plain Python by a factor of 2 to 120, achieves speeds comparable to that of C++, and often exceeds the speed of the existing solutions. We discuss the differences between HOPE and the other frameworks, as well as future extensions of its capabilities. The fully documented HOPE package is available at http://hope.phys.ethz.ch and is published under the GPLv3 license on PyPI and GitHub.

astro-ph.IM↗

arXiv2014

PynPoint Code for Exoplanet Imaging

We announce the public release of PynPoint, a Python package that we have developed for analysing exoplanet data taken with the angular differential imaging observing technique. In particular, PynPoint is designed to model the point spread function of the central star and to subtract its flux contribution to reveal nearby faint companion planets. The current version of the package does this correction by using a principal component analysis method to build a basis set for modelling the point spread function of the observations. We demonstrate the performance of the package by reanalysing publicly available data on the exoplanet beta Pictoris b, which consists of close to 24,000 individual image frames. We show that PynPoint is able to analyse this typical data in roughly 1.5 minutes on a Mac Pro, when the number of images is reduced by co-adding in sets of 5. The main computational work parallelises well as a result of a reliance on SciPy and NumPy functions. For this calculation the peak memory load is 6Gb, which can be run comfortably on most workstations. A simpler calculation, by co-adding over 50, takes 3 seconds with a peak memory usage of 600 Mb. This can be performed easily on a laptop. In developing the package we have modularised the code so that we will be able to extend functionality in future releases, through the inclusion of more modules, without it affecting the users application programming interface. We distribute the PynPoint package through the central PyPi sever, and the documentation is available online (http://pynpoint.ethz.ch).

astro-ph.EP↗