US 11,875,240 B1
Tuning a generative artificial intelligence model
Dusan Bosnjakovic, Los Angeles, CA (US); and Anshuman Sahu, Los Gatos, CA (US)
Assigned to Intuit Inc., Mountain View, CA (US)
Filed by Intuit Inc., Mountain View, CA (US)
Filed on Jul. 25, 2023, as Appl. No. 18/225,985.
Int. Cl. G06N 20/00 (2019.01); G06N 5/04 (2023.01); G06F 40/40 (2020.01)
CPC G06N 20/00 (2019.01) [G06F 40/40 (2020.01); G06N 5/04 (2013.01)] 20 Claims
OG exemplary drawing
 
1. A computer-implemented method for tuning a generative artificial intelligence (AI) model, the method comprising:
for each article of one or more articles stored in a knowledge base:
retrieving the article from the knowledge base;
receiving, by a natural language processing (NLP) model, the article retrieved from the knowledge base;
generating, by the NLP model, a question based on the article;
generating, by the NLP model, an answer based on the question, wherein the question and the answer are together a question answer pair;
receiving the question by a content retrieval model;
retrieving, by the content retrieval model, a second article from the knowledge base based on the question, wherein the second article is linked to the question answer pair; and
storing the question answer pair and the linked second article as a data point in a tuning data set; and
tuning a pretrained generative AI model using the tuning data set.