US 12,282,727 B2
Automatic document sketching
William B. Dolan, Kirkland, WA (US); Zeqiu Wu, Seattle, WA (US); Michel Galley, Seattle, WA (US); Yizhe Zhang, Seattle, WA (US); Zhang Li, Seattle, WA (US); and Christopher John Brockett, Bellevue, WA (US)
Assigned to Microsoft Technology Licensing, LLC, Redmond, WA (US)
Filed by Microsoft Technology Licensing, LLC, Redmond, WA (US)
Filed on Jan. 11, 2021, as Appl. No. 17/146,133.
Claims priority of provisional application 63/117,377, filed on Nov. 23, 2020.
Prior Publication US 2022/0164520 A1, May 26, 2022
Int. Cl. G06F 40/106 (2020.01); G06F 40/18 (2020.01); G06N 20/00 (2019.01)
CPC G06F 40/106 (2020.01) [G06F 40/18 (2020.01); G06N 20/00 (2019.01)] 20 Claims
OG exemplary drawing
 
1. A method of generating a document, the method comprising:
receiving an indication to create a document sketch;
determining, based on the indication, a plurality of example documents and grounding content;
receiving the plurality of example documents;
receiving the grounding content; and
generating, using a generative machine learning model, the document sketch based on the example documents and the grounding content, wherein the document sketch includes a plurality of generative text sequences that were produced by the generative machine learning model based at least in part on the grounding content, wherein the grounding content comprises relevant content for the document sketch and an identifier associated with the grounding content used to generate a predicted text sequence, and wherein grounding information derived from the grounding content is provided to the generative machine learning model to generate the predicted text sequence.