Granted Patent B2

US 12,499,141 · App. 17/472,250 · Granted Dec 16, 2025

Inventors: Yara Rizk (Cambridge, MA); Zilu Tang (Cambridge, MA); Yasaman Khazaeni (Needham, MA); Vatche Isahagian (Belmont, MA)

Assignee: International Business Machines Corporation

G06F16/3344G06F16/243G06F16/248G06F16/338G06F16/90332G06F40/169G06F40/284G06F40/30G06T11/206

Abstract

Claims (55)

1 . A computer-implemented method, comprising:

identifying an intent from terminology of a natural language query input by a user in a conversation within a chatbot interface of a conversational system;

receiving a natural language request from the user in the conversation within the conversational system for a data visualization;

utilizing verbiage from the natural language query, the identified intent, and entities identified from the natural language request to determine one or more data columns for the data visualization;

extracting, from the determined one or more data columns, one or more features indicative of a data-type of the determined one or more data columns;

annotating, based on an ontology, the extracted one or more features into a plot-label;

generating, based on the determined one or more data columns, multiple plots, wherein each plot of the generated multiple plots is of a different plot-type;

calculating, based on an amount of information associated with each plot of the generated multiple plots, a respective interestingness score for each plot of the generated multiple plots;

selecting a first plot of the generated multiple plots based on the respective interestingness score of the first plot being highest among interestingness scores of the generated multiple plots, and a plot-type of the first plot corresponding to the annotated plot-label; and

controlling display, in the conversation with the conversational system, of an ontology-based data visualization in a form of a graph that corresponds to the selected first plot, wherein the ontology-based data visualization leverages context of the conversation.

2 . The computer-implemented method of claim 1 , wherein the respective interestingness score of the first plot, that is highest, is less than a predefined threshold value.

3 . The computer-implemented method of claim 1 , further comprising:

requesting a feedback return for a quality of the ontology-based data visualization; and

learning from the feedback return with a trained supervised learning model by creating a threshold for an acceptable value for the highest interestingness score of the first plot based upon the trained supervised learning model trained at least in-part with user feedback.

4 . The computer-implemented method of claim 1 , further comprising:

requesting a feedback return for a quality of the ontology-based data visualization; and

learning from the feedback return by creating a threshold for an acceptable value for the highest interestingness score of the first plot and dynamically adjusting the ontology to change the annotated plot-label.

5 . The computer-implemented method of claim 1 , wherein the computer-implemented method is embodied in a cloud-computing environment.

6 . A computer program product comprising a computer-readable storage medium having program instructions embodied therewith, the program instructions executable by a computer to cause the computer to perform:

identifying an intent from terminology of a natural language query input by a user in a conversation within a chatbot interface of a conversational system;

receiving a natural language request from the user in the conversation within the conversational system for a data visualization;

utilizing verbiage from the natural language query, the identified intent, and entities identified from the natural language request to determine one or more data columns for the data visualization;

extracting, from the determined one or more data columns, one or more features indicative of a data-type of the determined one or more data columns;

annotating, based on an ontology, the extracted one or more features into a plot-label;

generating, based on the determined one or more data columns, multiple plots, wherein each plot of the generated multiple plots is of a different plot-type;

calculating, based on an amount of information associated with each plot of the generated multiple plots, a respective interestingness score for each plot of the generated multiple plots;

selecting a first plot of the generated multiple plots based on the respective interestingness score of the first plot being highest among interestingness scores of the generated multiple plots, and a plot-type of the first plot corresponding to the annotated plot-label; and

controlling display, in the conversation with the conversational system, of an ontology-based data visualization in a form of a graph that corresponds to the selected first plot, wherein the ontology-based data visualization leverages context of the conversation.

7 . The computer program product of claim 6 , wherein the respective interestingness score of the first plot, that is highest, is less than a predefined threshold value.

8 . The computer program product of claim 6 , wherein the program instructions further cause the computer to perform:

requesting a feedback return for a quality of the ontology-based data visualization; and

learning from the feedback return with a trained supervised learning model by creating a threshold for an acceptable value for the highest interestingness score of the first plot based upon the trained supervised learning model trained at least in-part with user feedback.

9 . The computer program product of claim 6 , wherein the program instructions further cause the computer to perform:

requesting a feedback return for a quality of the ontology-based data visualization; and

learning from the feedback return by creating a threshold for an acceptable value for the highest interestingness score of the first plot and dynamically adjusting the ontology to change the annotated plot-label.

10 . A system, comprising:

a processor; and

a memory, the memory storing instructions to cause the processor to:

identify an intent from terminology of a natural language query input by a user in a conversation within a chatbot interface of a conversational system;

receive a natural language request from the user in the conversation within the conversational system for a data visualization;

utilize verbiage from the natural language query, the identified intent, and entities identified from the natural language request to determine one or more data columns for the data visualization;

extract, from the determined one or more data columns, one or more features indicative of a data-type of the determined one or more data columns;

annotate, based on an ontology, the extracted one or more features into a plot-label;

generate, based on the determined one or more data columns, multiple plots, wherein each plot of the generated multiple plots is of a different plot-type;

calculate, based on an amount of information associated with each plot of the generated multiple plots, a respective interestingness score for each plot of the generated multiple plots;

select a first plot of the generated multiple plots based on the respective interestingness score of the first plot being highest among interestingness scores of the generated multiple plots, and a plot-type of the first plot corresponding to the annotated plot-label; and

control display, in the conversation with the conversational system, of an ontology-based data visualization in a form of a graph that corresponds to the selected first plot, wherein the ontology-based data visualization leverages context of the conversation.

11 . The system of claim 10 , wherein the respective interestingness score of the first plot, that is highest, is less than a predefined threshold value.

12 . The system of claim 10 , further cause the processor to:

request a feedback return for quality of the ontology-based data visualization; and

learn from the feedback return by creating a threshold for an acceptable value for the highest interestingness score of the first plot, and dynamically adjusting the ontology to change the annotated plot-label.

13 . The system of claim 10 , further cause the processor to:

request a feedback return with a trained supervised learning model for a quality of the ontology-based data visualization; and

learn from the feedback return by creating a threshold for an acceptable value for the highest interestingness score of the first plot based upon the trained supervised learning model trained at least in-part with user feedback.

14 . The system of claim 10 , wherein the system is embodied in a cloud-computing environment.

Assignments (1)

Continuity (1)

References Cited (28)