Subscribe to the gold package and get unlimited access to Shamra Academy
Register a new userMachine learning application to Fermi-LAT data: sharpening all-sky map and emphasizing variable sources
111
0
0.0
(
0
)
Ask ChatGPT about the research
No Arabic abstract
A novel application of machine-learning (ML) based image processing algorithms is proposed to analyze an all-sky map (ASM) obtained using the Fermi Gamma-ray Space Telescope. An attempt was made to simulate a one-year ASM from a short-exposure ASM generated from one-week observation by applying three ML based image processing algorithms: dictionary learning, U-net, and Noise2Noise. Although the inference based on ML is less clear compared to standard likelihood analysis, the quality of the ASM was generally improved. In particular, the complicated diffuse emission associated with the galactic plane was successfully reproduced only from one-week observation data to mimic a ground truth (GT) generated from a one-year observation. Such ML algorithms can be implemented relatively easily to provide sharper images without various assumptions of emission models. In contrast, large deviations between simulated ML maps and GT map were found, which are attributed to the significant temporal variability of blazar-type active galactic nuclei (AGNs) over a year. Thus, the proposed ML methods are viable not only to improve the image quality of an ASM, but also to detect variable sources, such as AGNs, algorithmically, i.e., without human bias. Moreover, we argue that this approach is widely applicable to ASMs obtained by various other missions; thus, it has the potential to examine giant structures and transient events, both of which are rarely found in pointing observations.
rate research
Read More
Classification of sources is one of the most important tasks in astronomy. Sources detected in one wavelength band, for example using gamma rays, may have several possible associations in other wavebands or there may be no plausible association candidates. In this work, we aim to determine probabilistic classification of unassociated sources in the third and the fourth data release 2 Fermi Large Area Telescope (LAT) point source catalogs (3FGL and 4FGL-DR2) into two classes (pulsars and active galactic nuclei (AGNs)) or three classes (pulsars, AGNs, and other sources). We use several machine learning (ML) methods to determine probabilistic classification of Fermi-LAT sources. We evaluate the dependence of results on meta-parameters of the ML methods, such as the maximal depth of the trees in tree-based classification methods and the number of neurons in neural networks. We determine probabilistic classification of both associated and unassociated sources in 3FGL and 4FGL-DR2 catalogs. We cross-check the accuracy by comparing the predicted classes of unassociated sources in 3FGL that have associations in 4FGL-DR2. We find that in the 2-class case it is important to correct for the presence of other sources among the unassociated ones in order to realistically estimate the number of pulsars and AGNs. In particular, the estimated number of pulsars in the 3FGL (4FGL-DR2) catalog is 270 (483) in the 2-class case without corrections for the other sources and 158 (215) in the 3-class case. Provided that the number of associated pulsars is 167 (271) in the 3FGL (4FGL-DR2) catalog, the number of pulsars among the unassociated sources is expected to be similar or larger than the number of associated ones.
Since 2008 August the Fermi Large Area Telescope (LAT) has provided continuous coverage of the gamma-ray sky yielding more than 5000 gamma-ray sources, but 54% of the detected sources remain with no certain or unknown association with a low energy counterpart. Rigorous determination of class type for a gamma-ray source requires the optical spectrum of the correct counterpart but optical observations are demanding and time-consuming, then machine learning techniques can be a powerful alternative for screening and ranking. We use machine learning techniques to select blazar candidates among uncertain sources characterized by gamma-ray properties very similar to those of Active Galactic Nuclei. Consequently, the percentage of sources of uncertain type drops from 54% to less than 12% predicting a new zoo for the Fermi gamma-ray sources. The result of this study opens up new considerations on the population of the gamma energy sky, and it will facilitate the planning of significant samples for rigorous analysis and multi-wavelength observational campaigns.
We present the first Fermi Large Area Telescope (LAT) low energy catalog (1FLE) of sources detected in the energy range 30 - 100 MeV. The COMPTEL telescope detected sources below 30 MeV, while catalogs released by the Fermi-LAT and EGRET collaborations use energies above 100 MeV. We create a list of sources detected in the energy range between 30 and 100 MeV, which closes a gap of point source analysis between the COMPTEL catalog and the Fermi-LAT catalogs. One of the main challenges in the analysis of point sources is the construction of the background diffuse emission model. In our analysis, we use a background-independent method to search for point-like sources based on a wavelet transform implemented in the PGWave code. The 1FLE contains 198 sources detected above 3 $sigma$ significance with eight years and nine months of the Fermi-LAT data. For 187 sources in the 1FLE catalog we have found an association in the Fermi-LAT 3FGL catalog: 148 are extragalactic, 22 are Galactic, and 17 are unclassified in the 3FGL. The ratio of the number of flat spectrum radio quasars (FSRQ) to BL Lacertae (BL Lacs) in 1FLE is 3 to 1, which can be compared with an approximately 1 to 1 ratio for the 3FGL or a 1 to 6 ratio for 3FHL. The higher ratio of the FSRQs in the 1FLE is expected due to generally softer spectra of FSRQs relative to BL Lacs. Most BL Lacs in 1FLE are of low-synchrotron peaked blazar type (18 out of 31), which have softer spectra and higher redshifts than BL Lacs on average. Correspondingly, we find that the average redshift of the BL Lacs in 1FLE is higher than in 3FGL or 3FHL. There are 11 sources that do not have associations in the 3FGL. Most of the unassociated sources either come from regions of bright diffuse emission or have several known 3FGL sources in the vicinity, which can lead to source confusion. The remaining unassociated sources have significance less than 4 $sigma$.
Four years into the mission, the understanding of the performance of the Fermi Large Area Telescope (LAT) and data analysis have increased enormously since launch. Thanks to a careful analysis of flight data, we were able to trace back some of the most significant sources of systematic uncertainties to using non-optimal calibration constants for some of the detectors. In this paper we report on a major effort within the LAT Collaboration to update these constants, to use them to reprocess the first four years of raw data, and to investigate the improvements observed for low- and high-level analysis. The Pass 7 reprocessed data, also known as P7REP data, are still being validated against the original Pass~7 (P7) data by the LAT Collaboration and should be made public, along with the corresponding instrument response functions, in the spring of 2013.
The analysis of Fermi Large Area Telescope (LAT) gamma-ray data in a given Region Of Interest (RoI) usually consists of performing a binned log-likelihood fit in order to determine the sky model that, after convolution with the instrument response, best accounts for the distribution of observed counts. While tools are available to perform such a fit, it is not easy to check the goodness-of-fit. The difficulty of the assessment of the data/model agreement is twofold. First of all, the observed and predicted counts are binned in three dimensions (two spatial dimensions and one energy dimension) and comparing two 3D maps is not straightforward. Secondly, gamma-ray source spectra generally decrease with energy as the inverse of the energy square. As a consequence the number of counts above several GeV generally falls into the Poisson regime, which precludes performing a simple $chi^2$ test. We propose a method that overcomes these two obstacles by producing and comparing spatially integrated count spectra for data and model at each pixel of the analysed RoI. The comparison is performed following a log-likelihood approach that extends the $chi^2$ test to histograms with low statistics. This method can take into account likelihood weights that are used to account for systematic uncertainties. We optimize the new method so that it provides a fast and reliable tool to assess the goodness-of-fit of Fermi-LAT data and we use it to check the latest gamma-ray source catalog on 10~years of data.
Log in to be able to interact and post comments
comments
Fetching comments
Sign in to be able to follow your search criteria
