papers in adversarial machine learning

A few weeks ago, Meta AI announced Galactica, a large language model (LLM) built for scientific work. Just for fun I asked it to write a blog post about adversarial machine learning. Galactica doesn't get anything obviously wrong, but repeats itself a lot, is fairly light on details, and makes tautological arguments.

Self-driving cars rely on vision for safety-critical information like traffic rules, which makes them susceptible to adversarial machine learning attacks. Some carefully placed stickers on a stop sign can make it invisible to autonomous vehicles; or, an adversarial t-shirt can make a person look like a stop sign.

What if breaking into an office was as easy as wearing a special pair of glasses, or putting a sticker on your forehead? It can be, if you make the right adversarial patch. Learn how to use adversarial machine learning to hide from face recognition systems, or convince them that you are someone else.

Sometimes, you need to send encrypted information, but also keep the fact that you are sending it a secret. Hiding secrets in regular data like this is called steganography, and it's cooler than it sounds, unless you are super into stegosaurus, and then it is exactly as cool as it sounds. With a few tweaks, you can use adversarial watermarking to hide information in normal-looking images and text. See how to do it here.

Machine learning models surpass human performance on image recognition tasks, but they can fail in surprising ways. By cataloguing these "natural" adversarial examples, you can learn a lot about how computer vision models work. You also learn that if you paint enough things yellow, computers will think the world is bananas.