US 12,266,345 B2
Automatic speech generation and intelligent and robust bias detection in automatic speech recognition model
Anup Bera, Greater Noida West (IN); and Hemant Palivela, Maharashtra (IN)
Assigned to Accenture Global Solutions Limited, Dublin (IE)
Filed by Accenture Global Solutions Limited, Dublin (IE)
Filed on Aug. 25, 2022, as Appl. No. 17/895,400.
Prior Publication US 2024/0071367 A1, Feb. 29, 2024
Int. Cl. G10L 15/00 (2013.01); G06F 3/0484 (2022.01); G10L 13/047 (2013.01); G10L 13/08 (2013.01); G10L 15/01 (2013.01); G10L 15/02 (2006.01); G10L 15/06 (2013.01); G10L 15/22 (2006.01); G10L 21/0216 (2013.01)
CPC G10L 15/01 (2013.01) [G06F 3/0484 (2013.01); G10L 13/047 (2013.01); G10L 13/08 (2013.01); G10L 15/02 (2013.01); G10L 15/063 (2013.01); G10L 15/22 (2013.01); G10L 21/0216 (2013.01); G10L 2015/025 (2013.01)] 20 Claims
OG exemplary drawing
 
1. A system comprising:
a memory circuitry for storing computer instructions;
a network interface circuitry; and
a processor in communication with the network interface circuitry and the memory circuitry, the processor configured to execute the computer instructions from the memory circuitry to:
receive speech samples uttered by a plurality of speakers;
determine a reference textual passage;
convert the reference textual passage into a set of machine-generated speeches corresponding to the plurality of speakers by automatically processing the reference textual passage and the speech samples using an automatic neural voice cloning model;
process the set of machine-generated speeches to produce a set of transcribed texts using at least one Automatic Speech Recognition (ASR) model; and
automatically quantify a bias in the at least one ASR model based on the set of transcribed texts and the reference textual passage.