MainMarket in brief


Researchers from Facebook and the University of California San Diego named NtechLab's deepfake video recognition algorithm the most resistant to attacks

December 15, 2020|

The Moscow-based NtechLab algorithm for deepfake video recognition was recognized by experts from the Facebook AI laboratory and the University of California at San Diego as the most resistant to deception using "adversarial attacks", subtle video distortions that allow you to deceive a neural network.

The researchers analysed the algorithm after it was placed third in the international Deepfake Detection Challenge in June this year.

Original (in Russian)








Interested In The Russian IT Market? Subscribe!

By clicking the button you agree to Privacy Policy

Unless otherwise stated, the content is available under Creative Commons BY 4.0 license

Supported by the Moscow Government