Hacking systems, like adversarial machine learning, can often teach us about how they work. In this video, you'll learn about the field of adversarial machine learning and how it relates to what you've learned thus far in memorization in AI systems.
Related Blog Post: https://blog.kjamistan.com/adversarial-examples-demonstrate-memorization-properties.html
My CCC talk (2017): https://media.ccc.de/v/34c3-8860-deep_learning_blindspots
Chart from: http://blog.datumbox.com/tuning-the-learning-rate-in-gradient-descent/
Towards Deep Learning Models Resistant to Adversarial Attacks (Mądry et al., 2018): https://arxiv.org/abs/1706.06083
Carlini et al. paper on diffusion fixes: https://arxiv.org/abs/2206.10550
I'm excited to hear what you liked from the videos thus far and what you want to learn about next, so please let me know in the comments!