Multi-frequency point source detection with fully convolutional networks: Performance in realistic microwave sky simulations

Abstract

Context. Point source (PS) detection is an important issue for future cosmic microwave background (CMB) experiments since they are one of the main contaminants to the recovery of CMB signal on small scales. Improving its multi-frequency detection would allow us to take into account valuable information otherwise neglected when extracting PS using a channel-by-channel approach. Aims. We aim to develop an artificial intelligence method based on fully convolutional neural networks to detect PS in multi-frequency realistic simulations and compare its performance against one of the most popular multi-frequency PS detection methods, the matrix filters. The frequencies used in our analysis are 143, 217, and 353 GHz, and we imposed a Galactic cut of 30 • . Methods. We produced multi-frequency realistic simulations of the sky by adding contaminating signals to the PS maps as the CMB, the cosmic infrared background, the Galactic thermal emission, the thermal Sunyaev-Zel'dovich effect, and the instrumental and PS shot noises. These simulations were used to train two neural networks called flat and spectral MultiPoSeIDoNs. The first one considers PS with a flat spectrum, and the second one is more realistic and general because it takes into account the spectral behaviour of the PS. Then, we compared the performance on reliability, completeness, and flux density estimation accuracy for both MultiPoSeIDoNs and the matrix filters. Results. Using a flux detection limit of 60 mJy, MultiPoSeIDoN successfully recovered PS reaching the 90% completeness level at 58 mJy for the flat case, and at 79, 71, and 60 mJy for the spectral case at 143, 217, and 353 GHz, respectively. The matrix filters reach the 90% completeness level at 84, 79, and 123 mJy. To reduce the number of spurious sources, we used a safer 4σ flux density detection limit for the matrix filters, the same as was used in the Planck catalogues, obtaining the 90% of completeness level at 113, 92, and 398 mJy. In all cases, MultiPoSeIDoN obtains a much lower number of spurious sources with respect to the filtering method. The recovering of the flux density of the detections, attending to the results on photometry, is better for the neural networks, which have a relative error of 10% above 100 mJy for the three frequencies, while the filter obtains a 10% relative error above 150 mJy for 143 and 217 GHz, and above 200 mJy for 353 GHz. Conclusions. Based on the results, neural networks are the perfect candidates to substitute filtering methods to detect multi-frequency PS in future CMB experiments. Moreover, we show that a multi-frequency approach can detect sources with higher accuracy than single-frequency approaches also based on neural networks.

Constraining the polarisation flux density and angle of point sources by training a convolutional neural network

“…However, when considering the application of this method in conjunction with other instruments, in order to obtain a good performance, the network should be retrained with a training set consisting of realistic simulations that reproduce the characteristics and conditions of such instruments, such as the higher resolution ACTPol or SPTPol experiments. Moreover, this method can be further improved in order to detect sources in polarisation data in a blind way by subsequently applying the fully convolutional neural network by Casas et al (2022b) to perform the detection in total intensity and then the methodology presented in this work to estimate the polarisation flux density and angle of the detected sources.…”

“…New methods based on neural networks and other ML methods have recently been developed in the field of CMB research, with promising results being obtained in both total intensity and polarisation. For example, Bonavera et al (2021) and Casas et al (2022b) compared their fully convolutional neural networks to commonly used filters in Planck for PS detection in singlefrequency (González-Nuevo et al 2006) and multi-frequency (Herranz et al 2009) realistic total-intensity simulations, obtain-ing more reliable results, and also at frequencies not used for training the networks.…”

CENN: A fully convolutional neural network for CMB recovery in realistic microwave sky simulations

“…They can optimise models with non-linear behaviours. Some recent applications in cosmology were in the PS detection field, both in single-frequency (Bonavera et al 2021) and in multi-frequency (Casas, J. M. et al 2022) approaches. Moreover, they have been used in the statistical study of the CMB for extending foreground models to subdegree angular scales (Krachmalnicoff & Puglisi 2021) and to perform a foreground model-recognition for B-mode observations (Farsian et al 2020).…”

CENN: a fully convolutional neural network for CMB recovery in realistic microwave sky simulations

“…One of the most important ML methods are the artificial neural networks, a mathematical approaches inspired on neuroscience, that can optimize models with nonlinear behaviors. Some recent applications in cosmology were in the PS detection field, both in single-frequency (Bonavera et al 2021) and in multi-frequency (Casas, J. M. et al 2022) approaches. Moreover, they have been used in the statistical study of the CMB for extending foreground models to sub-degree angular scales (Krachmalnicoff & Puglisi 2021) and to perform a foreground model recognition for B-mode observations (Farsian et al 2020).…”

“…The typical application of FCN is in the Euclidean space, i.e. to flat images as we did in Bonavera et al 2021 andin Casas, J. M. et al 2022. This is why our first step is to extract square patches from the full sky maps and work with such images when reconstructing the CMB.…”