It's a neat concept of passive modulation but looks easy to defeat with simple RF countermeasures. The 'Deep Learned' AI classifier is a powerful tool but as we are coming to realize with real-world applications it's easy to fool with simple adverse patterns. You can break it by slightly changing the problem in ways that humans find hard to detect but generate a completely different computer solution much like how optical illusions fool our perception of objects. It's like engineering bridges before physics, we still have no idea what's really happening with human intelligence but we can make massive structures that look impressive by adding pillars or a million pictures of RF patterns.At about the 1:00 minute mark it looks like one wireframe just kicked another wireframe's ass.
Not quite:We’re on more step closer to replicating every single one of Superman’s powers; We now have flying and x-ray vision at the same time. We just need to add the ability to be faster than a speeding train
by Jake Hertz
by Aaron Carman
by Aaron Carman