Black-box Generation of Adversarial Text Sequences to Evade Deep Learning Classifiers
Security researchers examine technical aspects of Black-box Generation of Adversarial Text Sequences to Evade Deep Learning Classifiers.
Summary
No summary yet — watch the video above, or open it on YouTube.
Related talks
IEEE Symposium on Security and Privacy·16 min
Data Hiding
Industry experts analyze emerging threats surrounding Adversarial Watermarking Transformer: Towards Tracing Text Provenance with Data Hiding.
IEEE Symposium on Security and Privacy·16 min
Engineers explore architecture and security mechanisms for AI-Guardian: Defeating Adversarial Attacks using Backdoors.
IEEE Symposium on Security and Privacy·15 min
Presenters detail vulnerability discovery and exploitation in REDEEM MYSELF: Purifying Backdoors in Deep Learning Models using Self Attention Distillation.