US 12,436,953 B2
Document generation system
Benjamin James Hoyle, Bath (GB); and Jerome Spaargaren, London (GB)
Assigned to Patently Limited, London (GB)
Filed by Patently Limited, London (GB)
Filed on Oct. 30, 2024, as Appl. No. 18/932,389.
Claims priority of provisional application 63/594,341, filed on Oct. 30, 2023.
Prior Publication US 2025/0139094 A1, May 1, 2025
Int. Cl. G06F 16/00 (2019.01); G06F 16/242 (2019.01); G06F 16/2455 (2019.01); G06F 40/284 (2020.01)
CPC G06F 16/2455 (2019.01) [G06F 16/242 (2019.01); G06F 40/284 (2020.01)] 23 Claims
OG exemplary drawing
 
1. A computer-implemented method of generating text associated with a figure, the method comprising:
receiving user-input unstructured text data describing one or more entities at least one of which is associated with the figure;
processing the unstructured text data to generate corresponding structured text data, and processing the structured text data to generate a set of parts;
determining a subset of the set of parts identifying parts that are present in the figure;
formulating prompt data for a neural network large language model to generate description text corresponding to the figure, the formulating of the prompt data comprising:
deriving description data associated with the figure from the user-input unstructured text data;
deriving part data corresponding to the subset of the set of parts for the figure; and
generating prompt data comprising the description data and the part data;
sending the prompt data to the neural network large language model; and
receiving description text data for the figure from the neural network large language model.