Root cause analysis of vulnerability of neural networks to adversarial examples Grant US-11443069-B2 United States of America 13 Sep 2022