{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/separating-the-eor-signal-with-a","title":"Separating the EoR Signal with a Convolutional Denoising Autoencoder: A Deep-learning-based Method","arxiv_id":"1902.09278","date":"2019-02-25","proceeding":null,"authors":["Weitian Li","Haiguang Xu","Zhixian Ma","Ruimin Zhu","Dan Hu","Zhenghao Zhu","Junhua Gu","Chenxi Shan","Jie Zhu","Xiang-Ping Wu"],"abstract":"When applying the foreground removal methods to uncover the faint\ncosmological signal from the epoch of reionization (EoR), the foreground\nspectra are assumed to be smooth. However, this assumption can be seriously\nviolated in practice since the unresolved or mis-subtracted foreground sources,\nwhich are further complicated by the frequency-dependent beam effects of\ninterferometers, will generate significant fluctuations along the frequency\ndimension. To address this issue, we propose a novel deep-learning-based method\nthat uses a 9-layer convolutional denoising autoencoder (CDAE) to separate the\nEoR signal. After being trained on the SKA images simulated with realistic beam\neffects, the CDAE achieves excellent performance as the mean correlation\ncoefficient ($\\bar{\\rho}$) between the reconstructed and input EoR signals\nreaches $0.929 \\pm 0.045$. In comparison, the two representative traditional\nmethods, namely the polynomial fitting method and the continuous wavelet\ntransform method, both have difficulties in modelling and removing the\nforeground emission complicated with the beam effects, yielding only\n$\\bar{\\rho}_{\\text{poly}} = 0.296 \\pm 0.121$ and $\\bar{\\rho}_{\\text{cwt}} =\n0.198 \\pm 0.160$, respectively. We conclude that, by hierarchically learning\nsophisticated features through multiple convolutional layers, the CDAE is a\npowerful tool that can be used to overcome the complicated beam effects and\naccurately separate the EoR signal. Our results also exhibit the great\npotential of deep-learning-based methods in future EoR experiments.","url_abs":"http://arxiv.org/abs/1902.09278v2","url_pdf":"http://arxiv.org/pdf/1902.09278v2.pdf","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[{"paper_slug":"separating-the-eor-signal-with-a","repo_url":"https://github.com/liweitianux/cdae-eor","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"none","reach":null}],"tasks":[{"task_slug":"denoising","task_name":"Denoising"}],"methods":[{"method_slug":"denoising-autoencoder","method_name":"Denoising Autoencoder"}],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"syntology_url":null,"atlas_url":null,"mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}