US 11,657,226 B2
Detecting and mitigating bias in natural language processing
Carla Quinn, Markham (CA); Keith Goode, Round Rock, TX (US); Mayo Takeuchi, Vienna (AT); and John Handy Bosma, Leander, TX (US)
Assigned to INTERNATIONAL BUSINESS MACHINES CORPORATION, Armonk, NY (US)
Filed by INTERNATIONAL BUSINESS MACHINES CORPORATION, Armonk, NY (US)
Filed on Dec. 15, 2020, as Appl. No. 17/121,789.
Prior Publication US 2022/0188518 A1, Jun. 16, 2022
Int. Cl. G06F 17/00 (2019.01); G06F 40/289 (2020.01); G06F 40/30 (2020.01); G06F 16/22 (2019.01)
CPC G06F 40/289 (2020.01) [G06F 16/2272 (2019.01); G06F 16/2291 (2019.01); G06F 40/30 (2020.01)] 17 Claims
OG exemplary drawing
 
1. A computer-based method of detecting and mitigating bias, the method comprising:
receiving real-world data from a database;
creating an inverted index from the real-world data;
analyzing words in the inverted index, wherein the analyzation identifies a plurality of categories in the real-world data;
generating a structure template containing various entities within each category of the plurality of categories;
receiving a test record of the structure template;
providing alternative entities in the test record where bias is likely to occur;
storing the test record;
providing an audit trail of the test record; and
in response to determining bias exists, indicating a corrective action.