While cool, I don’t think it’s accurate to say he implemented a version of FaceID. What he actually did was implement an algorithm that, for a limited case he tested, got results comparable to what FaceID would.
This isn’t about quality, it’s more a philosophical/empirical objection. We have no way of knowing if his approach is actually similar to Apple’s approach. They might do something entirely different, and it just happens that the results look similar to his results.
True, but it’s cool. I don’t know much about ML so the idea of a siamese network was totally novel to me. They didn’t go super in depth, but for me it was a delightful concept to be introduced to.
While cool, I don’t think it’s accurate to say he implemented a version of FaceID. What he actually did was implement an algorithm that, for a limited case he tested, got results comparable to what FaceID would.
This isn’t about quality, it’s more a philosophical/empirical objection. We have no way of knowing if his approach is actually similar to Apple’s approach. They might do something entirely different, and it just happens that the results look similar to his results.
True, but it’s cool. I don’t know much about ML so the idea of a siamese network was totally novel to me. They didn’t go super in depth, but for me it was a delightful concept to be introduced to.