On the problem of adversarial attacks in machine learning
A new paradigm is needed. But the moat around the current paradigm is so deep that I don't see how any scientist can carve out the time and space to pursue a different approach without committing career suicide.
No comments:
Post a Comment