Display options
Cite
Antun V, Renna F, Poon C, et al. On instabilities of deep learning in image reconstruction and the potential costs of AI. Proc Natl Acad Sci U S A. 2020;117(48):30088-30095doi: 10.1073/pnas.1907377117.
Antun, V., Renna, F., Poon, C., Adcock, B., & Hansen, A. C. (2020). On instabilities of deep learning in image reconstruction and the potential costs of AI. Proceedings of the National Academy of Sciences of the United States of America, 117(48), 30088-30095. https://doi.org/10.1073/pnas.1907377117
Antun, Vegard, et al. "On instabilities of deep learning in image reconstruction and the potential costs of AI." Proceedings of the National Academy of Sciences of the United States of America vol. 117,48 (2020): 30088-30095. doi: https://doi.org/10.1073/pnas.1907377117
Antun V, Renna F, Poon C, Adcock B, Hansen AC. On instabilities of deep learning in image reconstruction and the potential costs of AI. Proc Natl Acad Sci U S A. 2020 Dec 01;117(48):30088-30095. doi: 10.1073/pnas.1907377117. Epub 2020 May 11. PMID: 32393633; PMCID: PMC7720232.
Copy Download .nbib Format: NLM
Share it on

Proc Natl Acad Sci U S A. 2020 Dec 01;117(48):30088-30095. doi: 10.1073/pnas.1907377117. Epub 2020 May 11.

On instabilities of deep learning in image reconstruction and the potential costs of AI.

Proceedings of the National Academy of Sciences of the United States of America

Vegard Antun, Francesco Renna, Clarice Poon, Ben Adcock, Anders C Hansen

Affiliations

  1. Department of Mathematics, University of Oslo, 0316 Oslo, Norway.
  2. Instituto de Telecomunicações, Faculdade de Ciências, Universidade do Porto, Porto 4169-007, Portugal.
  3. Department of Mathematical Sciences, University of Bath, Bath BA2 7AY, United Kingdom.
  4. Department of Mathematics, Simon Fraser University, Burnaby, BC V5A 1S6, Canada.
  5. Department of Mathematics, University of Oslo, 0316 Oslo, Norway; [email protected].
  6. Department of Applied Mathematics and Theoretical Physics, University of Cambridge, Cambridge CB3 0WA, United Kingdom.

PMID: 32393633 PMCID: PMC7720232 DOI: 10.1073/pnas.1907377117

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.

Keywords: AI; deep learning; image reconstruction; instability; inverse problems

Conflict of interest statement

The authors declare no competing interest.

References

  1. Magn Reson Med. 2018 Jun;79(6):3055-3071 - PubMed
  2. IEEE Trans Med Imaging. 2018 Jun;37(6):1310-1321 - PubMed
  3. IEEE Trans Med Imaging. 2018 Jun;37(6):1440-1453 - PubMed
  4. Nature. 2015 May 28;521(7553):436-44 - PubMed
  5. Nature. 2018 Mar 21;555(7697):487-492 - PubMed
  6. IEEE Trans Image Process. 2017 Sep;26(9):4509-4522 - PubMed
  7. IEEE Trans Med Imaging. 2019 Jul;38(7):1633-1642 - PubMed
  8. Proc Natl Acad Sci U S A. 2012 Jun 26;109(26):E1679-87 - PubMed

Publication Types