# US 2019/0318000 A1 · Tool for Facilitating the Development of New Language Understanding Scenarios

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

- Number: US 2019/0318000 A1
- Document ID: US20190318000A1
- Status: Published application
- Kind: A1
- Application number: 16/451,994
- Filed: 2019-06-25
- Published: 2019-10-17
- Earliest claimed: 2016-01-23
- Assignees: Microsoft Technology Licensing, LLC
- Applicants: Microsoft Technology Licensing, LLC
- Inventors: Young-Bum Kim; Ruhi Sarikaya; Alexandre Rochette
- Web page: https://patentgrep.com/patent/US20190318000A1

## CPC

- G06F 40/40: Processing or translation of natural language (natural language analysis G06F40/20; semantic analysis G06F40/30)
- G06F 40/35: Discourse or dialogue representation
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)

## Abstract

A computer-implemented technique is described for facilitating the creation of a language understanding (LU) component for use with an application. The technique allows a developer to select a subset of parameters from a larger set of parameters. The subset of parameters pertains to a LU scenario to be handled by the application. The larger set of parameters pertains to a plurality of LU scenarios handled by an already-existing generic LU model. The technique creates a constrained LU component that is based on the subset of parameters in conjunction with the generic LU model. At runtime, the constrained LU component interprets input language items using the generic LU model in a manner that is constrained by the subset of parameters that have been selected, to provide an output result. The technique also allows the developer to create new rules and/or supplemental models.

## Claims

### Claim at position 1

1-20. (canceled)

### Claim 21

A system comprising:
one or more hardware processor devices; and
one or more computer-readable storage media storing machine-readable instructions which, when executed by the one or more hardware processor devices, cause the one or more hardware processor devices to:
access a language understanding model having a plurality of classifiers configured to output parameters reflecting results of classification operations performed by the classifiers for a plurality of language understanding scenarios;
receive input information specifying a subset of specific parameters selected from among the parameters of the language understanding model, the subset of specific parameters pertaining to a particular language understanding scenario to be handled by a particular application;
store customization logic identifying the subset of specific parameters in a data store; and
provide a constrained language understanding component based at least on the language understanding model, the constrained language understanding component being configured to perform interpretation of input language items based at least on the subset of specific parameters identified by the customization logic.

### Claim 22 (depends on 21)

The system of claim 21, wherein the machine-readable instructions, when executed by the one or more hardware processor devices, cause the one or more hardware processor devices to:
based at least on the customization logic, configure the constrained language understanding component to utilize a subset of classifiers selected from the plurality of classifiers of the language understanding model.

### Claim 23 (depends on 22)

The system of claim 22, wherein the machine-readable instructions, when executed by the one or more hardware processor devices, cause the one or more hardware processor devices to:
select the subset of classifiers based at least on the subset of specific parameters identified by the customization logic.

### Claim 24 (depends on 23)

The system of claim 23, wherein the subset of classifiers output the subset of specific parameters identified by the customization logic.

### Claim 25 (depends on 22)

The system of claim 22, wherein the machine-readable instructions, when executed by the one or more hardware processor devices, cause the one or more hardware processor devices to:
execute a development tool that is configured to apply the customization logic to the language understanding model and output the constrained language understanding component.

### Claim 26 (depends on 22)

The system of claim 22, wherein the language understanding model is trained based at least on a generic set of training examples to recognize various input language items associated with the plurality of language understanding scenarios, including at least one language understanding scenario other than the particular language understanding scenario to be handled by the particular application.

### Claim 27 (depends on 26)

The system of claim 26, wherein the plurality of classifiers of the language understanding model comprise a plurality of binary classifiers, and the subset of classifiers comprises individual binary classifiers selected from the plurality of binary classifiers of the language understanding model.

### Claim 28

A method implemented by one or more computing devices, the method comprising:
accessing a language understanding model having a plurality of classifiers configured to output parameters reflecting results of classification operations performed by the classifiers, the plurality of classifiers pertaining to a plurality of language understanding scenarios;
receiving input information specifying a subset of specific parameters selected from among the parameters of the language understanding model, the subset of specific parameters pertaining to a particular language understanding scenario;
storing customization logic identifying the subset of specific parameters in a data store; and
providing a constrained language understanding component based at least on the customization logic and the language understanding model, the constrained language understanding component being configured to perform interpretation of input language items for the particular language understanding scenario.

### Claim 29 (depends on 28)

The method of claim 28, wherein providing the constrained language understanding component comprises:
selectively reusing a subset of classifiers of the language understanding model in the constrained language understanding component, the subset of classifiers corresponding to the subset of specific parameters identified by the customization logic.

### Claim 30 (depends on 29)

The method of claim 29, wherein the selectively reusing comprises:
deactivating one or more other classifiers of the language understanding model.

### Claim 31 (depends on 29)

The method of claim 29, wherein the parameters output by the language understanding model comprise a plurality of intents pertaining to the plurality of language understanding scenarios, and the subset of specific parameters comprises a subset of intents pertaining to the particular language understanding scenario.

### Claim 32 (depends on 31)

The method of claim 31, wherein the language understanding model comprises a neural network model.

### Claim 33 (depends on 29)

The method of claim 29, wherein the parameters output by the language understanding model comprise a plurality of tags pertaining to the plurality of language understanding scenarios, and the subset of specific parameters comprises a subset of tags pertaining to the particular language understanding scenario.

### Claim 34 (depends on 33)

The method of claim 33, wherein the language understanding model comprises a Conditional Random Field model.

### Claim 35

A computer-readable storage medium storing computer-readable instructions which, when executed by one or more processor devices, cause the one or more processor devices to perform acts comprising:
accessing a language understanding model having a plurality of classifiers configured to output parameters reflecting results of classification operations performed by the classifiers;
receiving input information specifying a subset of specific parameters selected from among the parameters of the language understanding model, the subset of specific parameters pertaining to a language understanding scenario to be handled by an application;
storing customization logic reflecting the subset of specific parameters in a data store; and
based at least on the customization logic, customizing the language understanding model to obtain a constrained language understanding component that is configured to perform interpretation of input language items.

### Claim 36 (depends on 35)

The computer-readable storage medium of claim 35, the acts further comprising:
presenting a user interface that conveys multiple parameters output by the language understanding model; and
receiving the input information via the user interface.

### Claim 37 (depends on 36)

The computer-readable storage medium of claim 36, wherein the multiple parameters conveyed by the user interface comprise multiple intents, and the input information specifies at least one of the multiple intents.

### Claim 38 (depends on 36)

The computer-readable storage medium of claim 36, wherein the multiple parameters conveyed by the user interface comprise multiple tags, and the input information specifies at least one of the multiple tags.

### Claim 39 (depends on 35)

The computer-readable storage medium of claim 35, wherein the customizing further comprises:
selectively reusing certain parameters of the language understanding model and not other parameters of the language understanding model.

### Claim 40 (depends on 39)

The computer-readable storage medium of claim 39, wherein the selectively reusing further comprises:
deactivating certain classifiers of the language understanding model that output the other parameters.
