Computer engineers at Duke University have demonstrated that using complex numbers—numbers with both real and imaginary components—can play an integral part in securing artificial intelligence algorithms against malicious attacks that try to fool object-identifying software by subtly altering the images. By including just two complex-valued layers among hundreds if not thousands of training iterations, the technique can improve performance against such attacks without sacrificing any efficiency.
from News on Artificial Intelligence and Machine Learning https://ift.tt/3DDpT0v
Home
machine-learning-ai-news
News on Artificial Intelligence and Machine Learning
Imaginary numbers protect AI from very real threats
- Blogger Comment
- Facebook Comment
Subscribe to:
Post Comments
(
Atom
)
0 comments:
Post a Comment