Efficient adversarial debiasing with concept activation vector - Medical image case-studies.

Journal: Journal of biomedical informatics
Published Date: Jan 1, 2024

Abstract

BACKGROUND: A major hurdle for the real time deployment of the AI models is ensuring trustworthiness of these models for the unseen population. More often than not, these complex models are black boxes in which promising results are generated. However, when scrutinized, these models begin to reveal implicit biases during the decision making, particularly for the minority subgroups.

Authors

  • Ramon Correa
    School of Computing, Informatics, and Decision Systems Engineering, Arizona State University, Tempe, AZ, USA.
  • Khushbu Pahwa
    University of California Los Angeles, LA, USA.
  • Bhavik Patel
    Department of Radiology, Mayo Clinic, 5777 E Mayo Blvd, Phoenix, AZ, 85054, USA.
  • Celine M Vachon
  • Judy W Gichoya
    The Johns Hopkins Hospital, Department of Radiology, 601 N Caroline St, Room 4223, Baltimore, MD 21287 (S.K.); Cleveland Clinic, Department of Radiation Oncology, Cleveland, Ohio (H.E.); Emory University School of Medicine, Department of Radiology, Atlanta, Georgia (J.G.); University of Pennsylvania, Department of Radiology, Philadelphia, Pennsylvania (C.E.K.).
  • Imon Banerjee
    Mayo Clinic, Department of Radiology, Scottsdale, AZ, USA.

Keywords

External Resources

Popular Topics
Recent Journals