A team of computer scientists devised a way to quickly remove traces of sensitive user information from machine learning models.
  • poVoq
    1•3 years ago

    The article doesn’t really go into detail, but having a “synthetic dataset to negate” the personal data, sound to me like they still keep the personal data around (just somewhat obfuscated), but found a “clever” way to scrub it from the machine learning model’s output…