US 12,190,887 B2
Adversarial speech-text protection against automated analysis
Ngoc Minh Tran, Dublin (IE); Marco Simioni, Dublin (IE); and Hessel Tuinhof, Dublin (IE)
Assigned to International Business Machines Corporation, Armonk, NY (US)
Filed by International Business Machines Corporation, Armonk, NY (US)
Filed on Dec. 7, 2021, as Appl. No. 17/643,071.
Prior Publication US 2023/0178079 A1, Jun. 8, 2023
Int. Cl. G10L 15/26 (2006.01); G06N 20/00 (2019.01); G10L 21/0208 (2013.01)
CPC G10L 15/26 (2013.01) [G06N 20/00 (2019.01); G10L 21/0208 (2013.01)] 20 Claims
OG exemplary drawing
 
1. A method, comprising:
processing an audio signal comprising speech data;
transcribing the speech data to generate text data;
identifying a vulnerable portion of the text data;
in response to the identifying, modifying the text data to generate a robust transcript, wherein the modifying comprises replacing the vulnerable portion of the text data with adversarial text;
designing adversarial noise corresponding to the adversarial text; and
applying the corresponding adversarial noise to the audio signal to generate a robust audio signal comprising modified speech data that, when transcribed, generates a transcript with a similarity to the robust transcript that is above a threshold similarity.