	US 12,307,202 B2
	Automatic writing style detection and rewriting
Tomasz Lukasz Religa, Seattle, WA (US); Warren Aldred, Redmond, WA (US); Si-Qing Chen, Bellevue, WA (US); Zhang Li, Sammamish, WA (US); Jesse Alexander Freitas, Seattle, WA (US); Tao Ge, Beijing (CN); Huitian Jiao, Snoqualmie, WA (US); Max Wang, Seattle, WA (US); and Xun Wang, Amherst, MA (US)
Assigned to Microsoft Technology Licensing, LLC, Redmond, WA (US)
Filed by Microsoft Technology Licensing, LLC, Redmond, WA (US)
Filed on Aug. 18, 2022, as Appl. No. 17/890,714.
Prior Publication US 2024/0061999 A1, Feb. 22, 2024
Int. Cl. G06F 40/253 (2020.01); G06F 40/166 (2020.01)

CPC G06F 40/253 (2020.01) [G06F 40/166 (2020.01)]

20 Claims

1. A data processing system comprising:

a processor; and

a machine-readable storage medium storing executable instructions that, when executed, cause the processor to perform operations comprising:

obtaining source textual content that is associated with a first writing style;

obtaining first target textual content that is associated with a second writing style that is different from the first writing style;

providing the source textual content as an input to a first machine learning model trained to determine a writing style of textual content provided as the input;

analyzing the source textual content using the first machine learning model to obtain an indication of the writing style of the source textual content;

providing the indication of the writing style of the source textual content and the first target textual content as an input to a second machine learning model trained to analyze the first target textual content and to rewrite the first target textual content according to the writing style represented by the indication that was obtained from the first machine learning model;

analyzing the indication of the writing style of the source textual content and the first target textual content using the second machine learning model to obtain updated first target textual content that has been rewritten according to the first writing style;

storing writing style information for the first writing style in a writing style datastore;

obtaining a second target textual content that is associated with a third writing style;

receiving a request to apply the first writing style to the second target textual content;

retrieving the writing style information for the first writing style from the writing style datastore;

providing the writing style information for the first writing style and the second target textual content as an input to the second machine learning model; and

analyzing the writing style information for the first writing style and the second target textual content using the second machine learning model to obtain updated second target textual content that has been rewritten according to the first writing style.