Subscribe to the gold package and get unlimited access to Shamra Academy
Register a new userMachine Learning on Difference Image Analysis: A comparison of methods for transient detection
87
0
0.0
(
0
)
Ask ChatGPT about the research
No Arabic abstract
We present a comparison of several Difference Image Analysis (DIA) techniques, in combination with Machine Learning (ML) algorithms, applied to the identification of optical transients associated with gravitational wave events. Each technique is assessed based on the scoring metrics of Precision, Recall, and their harmonic mean F1, measured on the DIA results as standalone techniques, and also in the results after the application of ML algorithms, on transient source injections over simulated and real data. This simulations cover a wide range of instrumental configurations, as well as a variety of scenarios of observation conditions, by exploring a multi dimensional set of relevant parameters, allowing us to extract general conclusions related to the identification of transient astrophysical events. The newest subtraction techniques, and particularly the methodology published in Zackay et al. (2016) are implemented in an Open Source Python package, named properimage, suitable for many other astronomical image analyses. This together with the ML libraries we describe, provides an effective transient detection software pipeline. Here we study the effects of the different ML techniques, and the relative feature importances for classification of transient candidates, and propose an optimal combined strategy. This constitutes the basic elements of pipelines that could be applied in searches of electromagnetic counterparts to GW sources.
rate research
Read More
Among the many challenges posed by the huge data volumes produced by the new generation of astronomical instruments there is also the search for rare and peculiar objects. Unsupervised outlier detection algorithms may provide a viable solution. In this work we compare the performances of six methods: the Local Outlier Factor, Isolation Forest, k-means clustering, a measure of novelty, and both a normal and a convolutional autoencoder. These methods were applied to data extracted from SDSS stripe 82. After discussing the sensitivity of each method to its own set of hyperparameters, we combine the results from each method to rank the objects and produce a final list of outliers.
Efficient identification and follow-up of astronomical transients is hindered by the need for humans to manually select promising candidates from data streams that contain many false positives. These artefacts arise in the difference images that are produced by most major ground-based time domain surveys with large format CCD cameras. This dependence on humans to reject bogus detections is unsustainable for next generation all-sky surveys and significant effort is now being invested to solve the problem computationally. In this paper we explore a simple machine learning approach to real-bogus classification by constructing a training set from the image data of ~32000 real astrophysical transients and bogus detections from the Pan-STARRS1 Medium Deep Survey. We derive our feature representation from the pixel intensity values of a 20x20 pixel stamp around the centre of the candidates. This differs from previous work in that it works directly on the pixels rather than catalogued domain knowledge for feature design or selection. Three machine learning algorithms are trained (artificial neural networks, support vector machines and random forests) and their performances are tested on a held-out subset of 25% of the training data. We find the best results from the random forest classifier and demonstrate that by accepting a false positive rate of 1%, the classifier initially suggests a missed detection rate of around 10%. However we also find that a combination of bright star variability, nuclear transients and uncertainty in human labelling means that our best estimate of the missed detection rate is approximately 6%.
We present a method for characterizing image-subtracted objects based on shapelet analysis to identify transient events in ground-based time-domain surveys. We decompose the image-subtracted objects onto a set of discrete Zernike polynomials and use their resulting coefficients to compare them to other point-like objects. We derive a norm in this Zernike space that we use to score transients for their point-like nature and show that it is a powerful comparator for distinguishing image artifacts, or residuals, from true astrophysical transients. Our method allows for a fast and automated way of scanning overcrowded, wide-field telescope images with minimal human interaction and we reduce the large set of unresolved artifacts left unidentified in subtracted observational images. We evaluate the performance of our method using archival intermediate Palomar Transient Factory and Dark Energy Camera survey images. However, our technique allows flexible implementation for a variety of different instruments and data sets. This technique shows a reduction in image subtraction artifacts by 99.95% for surveys extending up to hundreds of square degrees and has strong potential for automated transient identification in electromagnetic follow-up programs triggered by the Laser Interferometer Gravitational Wave Observatory-Virgo Scientific Collaboration.
Every field of Science is undergoing unprecedented changes in the discovery process, and Astronomy has been a main player in this transition since the beginning. The ongoing and future large and complex multi-messenger sky surveys impose a wide exploiting of robust and efficient automated methods to classify the observed structures and to detect and characterize peculiar and unexpected sources. We performed a preliminary experiment on KiDS DR4 data, by applying to the problem of anomaly detection two different unsupervised machine learning algorithms, considered as potentially promising methods to detect peculiar sources, a Disentangled Convolutional Autoencoder and an Unsupervised Random Forest. The former method, working directly on images, is considered potentially able to identify peculiar objects like interacting galaxies and gravitational lenses. The latter instead, working on catalogue data, could identify objects with unusual values of magnitudes and colours, which in turn could indicate the presence of singularities.
We show that multiple machine learning algorithms can match human performance in classifying transient imaging data from the Sloan Digital Sky Survey (SDSS) supernova survey into real objects and artefacts. This is a first step in any transient science pipeline and is currently still done by humans, but future surveys such as the Large Synoptic Survey Telescope (LSST) will necessitate fully machine-enabled solutions. Using features trained from eigenimage analysis (principal component analysis, PCA) of single-epoch g, r and i-difference images, we can reach a completeness (recall) of 96 per cent, while only incorrectly classifying at most 18 per cent of artefacts as real objects, corresponding to a precision (purity) of 84 per cent. In general, random forests performed best, followed by the k-nearest neighbour and the SkyNet artificial neural net algorithms, compared to other methods such as naive Bayes and kernel support vector machine. Our results show that PCA-based machine learning can match human success levels and can naturally be extended by including multiple epochs of data, transient colours and host galaxy information which should allow for significant further improvements, especially at low signal-to-noise.
Log in to be able to interact and post comments
comments
Fetching comments
Sign in to be able to follow your search criteria
