Existing Model Ownership Resolution schemes are vulnerable to false claims by malicious actors, exploiting transferable adversarial examples.
