{"database":"biostudies-literature","file_versions":[],"scores":null,"additional":{"submitter":["Antun V"],"funding":["Canadian Network for Research and Innovation in Machining Technology, Natural Sciences and Engineering Research Council of Canada","Leverhulme Trust","Nvidia","Royal Society","European Commission","RCUK | Engineering and Physical Sciences Research Council","Engineering and Physical Sciences Research Council"],"pagination":["30088-30095"],"full_dataset_link":["https://www.ebi.ac.uk/biostudies/studies/S-EPMC7720232"],"repository":["biostudies-literature"],"omics_type":["Unknown"],"volume":["117(48)"],"pubmed_abstract":["Deep learning, due to its unprecedented success in tasks such as image classification, has emerged as a new tool in image reconstruction with potential to change the field. In this paper, we demonstrate a crucial phenomenon: Deep learning typically yields unstable methods for image reconstruction. The instabilities usually occur in several forms: 1) Certain tiny, almost undetectable perturbations, both in the image and sampling domain, may result in severe artefacts in the reconstruction; 2) a small structural change, for example, a tumor, may not be captured in the reconstructed image; and 3) (a counterintuitive type of instability) more samples may yield poorer performance. Our stability test with algorithms and easy-to-use software detects the instability phenomena. The test is aimed at researchers, to test their networks for instabilities, and for government agencies, such as the Food and Drug Administration (FDA), to secure safe use of deep learning methods."],"journal":["Proceedings of the National Academy of Sciences of the United States of America"],"pubmed_title":["On instabilities of deep learning in image reconstruction and the potential costs of AI."],"pmcid":["PMC7720232"],"funding_grant_id":["1","611675","EP/L003457/1","SFRH/BPD/118714/2016","655282"],"pubmed_authors":["Poon C","Antun V","Hansen AC","Renna F","Adcock B"],"additional_accession":[]},"is_claimable":false,"name":"On instabilities of deep learning in image reconstruction and the potential costs of AI.","description":"Deep learning, due to its unprecedented success in tasks such as image classification, has emerged as a new tool in image reconstruction with potential to change the field. In this paper, we demonstrate a crucial phenomenon: Deep learning typically yields unstable methods for image reconstruction. The instabilities usually occur in several forms: 1) Certain tiny, almost undetectable perturbations, both in the image and sampling domain, may result in severe artefacts in the reconstruction; 2) a small structural change, for example, a tumor, may not be captured in the reconstructed image; and 3) (a counterintuitive type of instability) more samples may yield poorer performance. Our stability test with algorithms and easy-to-use software detects the instability phenomena. The test is aimed at researchers, to test their networks for instabilities, and for government agencies, such as the Food and Drug Administration (FDA), to secure safe use of deep learning methods.","dates":{"release":"2020-01-01T00:00:00Z","publication":"2020 Dec","modification":"2024-11-20T01:05:24.677Z","creation":"2021-02-20T13:42:28Z"},"accession":"S-EPMC7720232","cross_references":{"pubmed":["32393633"],"doi":["10.1073/pnas.1907377117"]}}