How Much Food Coloring Can Your Robot Handle? An Intro to Poisoning Machine Learning Systems

Presented at HOPE 2020 Virtual Rescheduled, July 28, 2020, 10 p.m. (60 minutes).

Machine learning has lately hit the buzz word spotlight. Finding both practical and impractical applications in fields from neuroscience to information security to... ranking bachelor contestants? This talk will cover some of the basics of manipulating and evading machine learning systems of all kinds, including how to confuse some military or government surveillance systems. This will cover the most common and useful attacks, how to apply them, and how to defend against these in the future. This is useful both for testing systems, as well as understanding how to specify and build them.


Presenters:

  • Corbin Frisvold
    **Corbin Frisvold (@QuesoSec)** is a 16-year-old hacker and polymath, involved in many fields such as mathematics, physics, computer science, and electrical engineering. He attended his first infosec conference (BSidesDE) at six years old and hasn't looked back. Since then he has worked on many significant projects, including publishing research papers on evolutionary soft robotics, running a village for kids called SpawnCamp, and running a blog at maker.godshell.com. Currently he is working on projects involving machine learning and its applications to robotics.<br>

Links:

Similar Presentations: