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

• Type: Article in International Scientific Journal
• Date: 2020

Authors

Antun, V (Author) Other; Renna, F (Author) FCUP; Poon, C (Author) Other; Adcock, B (Author) Other; Hansen, AC (Author) Other

Journal

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

Vol. 117

ISSN: 0027-8424

Publisher: National Academy of Sciences

Indexing

ISI Web of Knowledge - 266 Citations

Other information

Authenticus ID: P-00T-754

DOI: 10.1073/pnas.1907377117

Abstract (EN): 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.

Language: English

Type (Professor's evaluation): Scientific

No. of pages: 8