# US 12,182,493 B2 · Facilitating generation and presentation of advanced insights

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 12,182,493 B2
- Document ID: US12182493B2
- Status: Granted
- Kind: B2
- Application number: 18/484,674
- Filed: 2023-10-11
- Published: 2024-03-21
- Granted: 2024-12-31
- Earliest claimed: 2022-09-21
- Assignees: Adobe Inc.
- Applicants: Adobe Inc.
- Inventors: Md Main Uddin Rony; Fan Du; Iftikhar Ahamath Burhanuddin; Ryan Rossi; Niyati Himanshu Chhaya; Eunyee Koh
- Web page: https://patentgrep.com/patent/US12182493B2

## CPC

- G06F 40/106: Display of layout of documents; Previewing
- G06F 40/30: Semantic analysis
- G06F 40/186: Templates
- G06F 40/40: Processing or translation of natural language (natural language analysis G06F40/20; semantic analysis G06F40/30)
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)

## Abstract

Methods, computer systems, computer-storage media, and graphical user interfaces are provided for facilitating generation and presentation of insights. In one implementation, a set of data is used to generate a data visualization. A candidate insight associated with the data visualization is generated, the candidate insight being generated in text form based on a text template and comprising a descriptive insight, a predictive insight, an investigative, or a prescriptive insight. A set of natural language insights is generated, via a machine learning model. The natural language insights represent the candidate insight in a text style that is different from the text template. A natural language insight having the text style corresponding with a desired text style is selected for presenting the candidate insight and, thereafter, the selected natural language insight and data visualization are providing for display via a graphical user interface.

## Claims

### Claim 1

A system comprising:
a memory component; and
a processing device coupled to the memory component, the processing device to perform operations comprising:
causing display, via a graphical user interface, of a data visualization generated in accordance with a set of data;
causing display, via the graphical user interface, of a natural language insight in an insight panel presented concurrently with the data visualization, wherein the natural language insight is selected for display by:
generating a candidate insight associated with the data visualization, the candidate insight being generated in text form based on a text template and comprising a descriptive insight, a predictive insight, an investigative insight, or a prescriptive insight;
generating, via a machine learning model, a plurality of natural language insights that each represent the candidate insight in a text style different from the text template; and
selecting the natural language insight for display based on a desired text style for presenting the candidate insight.

### Claim 2 (depends on 1)

The system of claim 1,
wherein if the candidate insight comprises the descriptive insight, the descriptive insight indicates information describing the data visualization,
wherein if the candidate insight comprises the predictive insight, the predictive insight predicts future data points,
wherein if the candidate insight comprises the investigative insight, the investigative insight indicates a rational for the descriptive insight, and
wherein if the candidate insight comprises the prescriptive insight, the prescriptive insight indicates a recommended action to perform based on the descriptive insight.

### Claim 3 (depends on 1)

The system of claim 1, wherein the desired text style is a user-selected text style input via the graphical user interface or is determined based on previous user interactions.

### Claim 4 (depends on 1)

The system of claim 1, further comprising determining a set of annotation features for visually annotating the natural language insight.

### Claim 5 (depends on 1)

The system of claim 1, wherein the machine learning model comprises a multilayer perceptron model and uses counterfactual reasoning to generate the set of natural language insights.

### Claim 6 (depends on 1)

The system of claim 1, wherein the plurality of natural language insights are generated by:
using a first encoder-decoder to encode the candidate insight in text form into a disentangled representation including a content representation and a style representation;
using the machine learning model, in the form of a counterfactual reasoning module, to generate a plurality of different style representations based on the style representation; and
using a second encoder-decoder to generate the plurality of natural language insights, wherein each of the natural language insights is generated by aggregating a corresponding different style representation with the content representation.

### Claim 7 (depends on 1)

The system of claim 1, wherein the natural language insight is selected based on corresponding confidences indicating an extent of text style similarity corresponding with the desired text style.

### Claim 8

A non-transitory computer-readable medium storing executable instructions, which when executed by a processing device, cause the processing device to perform operations comprising:
causing display, via a graphical user interface, of a data visualization generated using a set of data;
causing display, via the graphical user interface, of an insight panel presented concurrently with the data visualization, wherein the insight panel includes:
a descriptive insight indicating information describing the data visualization using at least a portion of the set of data; and
an investigative insight indicating a rational for the descriptive insight.

### Claim 9 (depends on 8)

The non-transitory computer-readable medium of claim 8, wherein the insight panel further includes a prescriptive insight indicating a recommended action to perform based on the descriptive insight.

### Claim 10 (depends on 9)

The non-transitory computer-readable medium of claim 9, wherein identified investigative insight or the prescriptive insight is presented based on a user selection associated with the descriptive insight.

### Claim 11 (depends on 8)

The non-transitory computer-readable medium of claim 8, wherein the descriptive insight comprises a cyclic pattern, an extremum data, or a dramatic change.

### Claim 12 (depends on 10)

The non-transitory computer-readable medium of claim 10, wherein a dramatic change is identified as a descriptive insight when a difference between a highest value and a lowest value for a given period is less than a percentage of that difference.

### Claim 13 (depends on 8)

The non-transitory computer-readable medium of claim 8, wherein the data visualization is presented within a first panel in the graphical user interface and the descriptive insight is presented within a second panel in the graphical user interface, and wherein the investigative insight is presented based on a user selection associated with the descriptive insight presented in the second panel in the graphical user interface.

### Claim 14 (depends on 8)

The non-transitory computer-readable medium of claim 8, wherein at least one of the descriptive insight or the investigative insight includes a text style selected to correspond to a desired text style for presenting the corresponding insight.

### Claim 15 (depends on 8)

The non-transitory computer-readable medium of claim 8, wherein the descriptive insight includes a visual annotation for visually annotating the descriptive insight in association with the data visualization.

### Claim 16

A system comprising:
a memory component; and
a processing device coupled to the memory component, the processing device to perform operations comprising:
causing display, via a graphical user interface, of a data visualization generated using a set of data;
causing display, via the graphical user interface, of a descriptive insight indicating information describing the data visualization, the descriptive insight being generated in text form using at least a portion of the set of data; and
causing display, via the graphical user interface, of a set of annotation features that visually depict the text form of the descriptive insight in association with the data visualization.

### Claim 17 (depends on 16)

The system of claim 16, wherein the set of annotation features used to visually depict the text form of the descriptive insight are presented in a manner that corresponds with an annotation attribute identified based on a user selection or a set of previous user interactions.

### Claim 18 (depends on 16)

The system of claim 16, wherein the set of annotation features includes a set of one or more points in a graph or a chart to visually emphasize.

### Claim 19 (depends on 16)

The system of claim 16, wherein the set of annotation features visually depict the text form of the descriptive insight in a new chart or a new graph presented in association with the descriptive insight.

### Claim 20 (depends on 16)

The system of claim 16, wherein the text form of the descriptive insight is a natural language insight generated, via a machine learning model, in a text style different from a text template.
