Granted Patent B2

US 12,608,350 · App. 18/337,891 · Granted Apr 21, 2026

Inventors: Bongshin Lee (Issaquah, WA); Chenglong Wang (Bellevue, WA); John Roger Thompson (Atlanta, GA)

Assignee: Microsoft Technology Licensing, LLC

G06F16/215G06N3/02G06N3/0475

Abstract

Claims (26)

1 . A computer-implemented method, executed on a computing device, comprising:

processing a request to generate a visualization concerning a plurality of data concepts, wherein each of the data concepts comprises related portions of data of a data set;

generating a new data concept by transforming an existing data concept from the plurality of data concepts using a program synthesizer and a generative machine learning model, wherein generating the new data concept includes deriving the new data concept from the existing data concept using the generative machine learning model by:

processing a natural language prompt concerning the new data concept using the generative machine learning model;

providing a plurality of candidate data transformations from the generative machine learning model based on at least the natural language prompt and an author interaction history;

generating a sample entry for each candidate data transformation of the plurality of candidate data transformations; and

providing a representation of a process for generating the sample entry of each candidate data transformation by providing a code segment used to generate the sample entry of each candidate data transformation of the plurality of candidate data transformations; and

generating a visualization of the new data concept by processing a mapping of the new data concept to a visual channel of the visualization of the new data concept for rendering.

2 . The computer-implemented method of claim 1 , wherein generating the new data concept for the visualization includes reshaping the existing data concept using one or more of the program synthesizer and the generative machine learning model.

3 . The computer-implemented method of claim 2 , wherein reshaping the existing data concept includes prompting a user to provide an example of the new data concept.

4 . The computer-implemented method of claim 3 , wherein reshaping the existing data concept includes generating a candidate data transformation by processing the example of the new data concept using one or more of the program synthesizer and the generative machine learning model.

5 . A computing system comprising:

a memory; and

a processor configured to process a request to generate a visualization concerning a plurality of data concepts in a table, wherein each of the data concepts comprises related portions of data of a data set, to generate a new data concept from the plurality of data concepts for the visualization by deriving the new data concept from an existing data concept using a generative machine learning model, wherein deriving the new data concept includes: processing a natural language prompt concerning the new data concept using the generative machine learning model, providing a plurality of candidate data transformations from the generative machine learning model in response to processing the natural language prompt, generating a sample entry for each candidate data transformation of the plurality of candidate data transformations, providing a representation of a process for generating the sample entry of each candidate data transformation by providing a code segment used to generate the sample entry of each candidate data transformation of the plurality of candidate data transformations, and generating the new data concept by processing a selection of a candidate data transformation to transform the existing data concept, and to generate a visualization of the new data concept by processing a mapping of the new data concept to a visual channel of the visualization of the new data concept for rendering.

6 . The computing system of claim 5 , wherein the generative machine learning model includes one or more of:

a transformer-based multimodal Large Language Model (LLM); or

a diffusion-based generative multimodal model.

7 . A storage device for storing a computer program product residing on a computer readable medium having a plurality of instructions stored thereon which, when executed by a processor, cause the processor to perform operations comprising:

processing a request to generate a visualization concerning a plurality of data concepts in a table, wherein each of the data concepts comprises related portions of data of a data set;

generating a new data concept for the visualization by reshaping an existing data concept from the plurality of data concepts using a program synthesizer and a generative machine learning model, wherein generating the new data concept includes deriving the new data concept from the existing data concept using the generative machine learning model by:

processing a natural language prompt concerning the new data concept using the generative machine learning model;

providing a plurality of candidate data transformations from the generative machine learning model based on at least the natural language prompt and an author interaction history;

generating a sample entry for each candidate data transformation of the plurality of candidate data transformations; and

providing a representation of a process for generating the sample entry of each candidate data transformation by providing a code segment used to generate the sample entry of each candidate data transformation of the plurality of candidate data transformations; and

generating a visualization of the new data concept by processing a mapping of the new data concept to a visual channel of the visualization of the new data concept for rendering.

8 . The computer program product of claim 7 , wherein the candidate data transformation includes a candidate data table.

Assignments (1)

Continuity (1)

References Cited (82)