# US 2021/0334469 A1 · Adapting dialog model for completing task in dialog system

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

- Number: US 2021/0334469 A1
- Document ID: US20210334469A1
- Status: Published application
- Kind: A1
- Application number: 16/860,819
- Filed: 2020-04-28
- Published: 2021-10-28
- Earliest claimed: 2020-04-28
- Assignees: International Business Machines Corporation
- Applicants: International Business Machines Corporation
- Inventors: Song Feng; Qingzi Liao; Luis A. Lastras-Montano; Robert G. Farrell; Ana Smith
- Web page: https://patentgrep.com/patent/US20210334469A1

## CPC

- G06F 40/30: Semantic analysis
- G06F 40/284: Lexical analysis, e.g. tokenisation or collocates
- G06N 5/02: Knowledge representation; Symbolic representation
- G06N 5/04: Inference or reasoning models
- G06N 20/00: Machine learning
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)
- G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS

## Abstract

A computer-implemented method identifies one or more concepts from a document, wherein the document comprises information on completing a task. The method computes a relevance value of an utterance with respect to completing the task using the one or more identified concepts. The method removes the utterance from a dialog model to be used for completing the task when the relevance value of the utterance is below a given threshold value.

## Claims

### Claim 1

An apparatus comprising:
at least one processor;
at least one memory including computer program code;
the at least one memory and the computer program code being configured to, with the at least one processor, cause the apparatus at least to:
identify one or more concepts from a document, wherein the document comprises information on completing a task;
compute a relevance value of an utterance with respect to completing the task using the one or more identified concepts; and
remove the utterance from a dialog model to be used for completing the task when the relevance value of the utterance is below a given threshold value.

### Claim 2 (depends on 1)

The apparatus of claim 1, wherein identifying the one or more concepts from the document comprises a graph-based decomposition process.

### Claim 3 (depends on 2)

The apparatus of claim 2, wherein computing the relevance value comprises determining a measure of centrality associated with a text graph.

### Claim 4 (depends on 1)

The apparatus of claim 1, wherein identifying the one or more concepts from the document further comprises computation of one or more of a term frequency, an inverse document frequency, and a term frequency-inverse document frequency.

### Claim 5 (depends on 1)

The apparatus of claim 1, wherein the one or more concepts comprise one or more of linguistic constituents and tokens.

### Claim 6 (depends on 1)

The apparatus of claim 1, wherein computing the relevance value further comprises computing one or more of a maximum marginal relevance value, an integer linear programming value, and a sequential minimal optimization value.

### Claim 7 (depends on 1)

The apparatus of claim 1, wherein when the relevance value of the utterance is below the given threshold value, the utterance is considered a non-essential utterance with respect to completing the task and is removed from the dialog model.

### Claim 8 (depends on 7)

The apparatus of claim 7, wherein when the relevance value of the utterance is not below the given threshold value, the utterance is considered an essential utterance with respect to completing the task and is included in the dialog model.

### Claim 9 (depends on 1)

The apparatus of claim 1, wherein the relevance value is a measure of whether or not the utterance contains information relevant to completing the task.

### Claim 10

A computer-implemented method comprising:
identifying one or more concepts from a document, wherein the document comprises information on completing a task;
computing a relevance value of an utterance with respect to completing the task using the one or more identified concepts; and
removing the utterance from a dialog model to be used for completing the task when the relevance value of the utterance is below a given threshold value.

### Claim 11 (depends on 10)

The method of claim 10, wherein identifying the one or more concepts from the document further comprises a graph-based decomposition process.

### Claim 12 (depends on 11)

The method of claim 11, wherein computing the relevance value comprises determining a measure of centrality associated with a text graph.

### Claim 13 (depends on 10)

The method of claim 10, wherein identifying the one or more concepts from the document further comprises computation of one or more of a term frequency, an inverse document frequency, and a term frequency-inverse document frequency.

### Claim 14 (depends on 10)

The method of claim 10, wherein the one or more concepts comprise one or more of linguistic constituents and tokens.

### Claim 15 (depends on 10)

The method of claim 10, wherein computing the relevance value further comprises computing one or more of a maximum marginal relevance value, an integer linear programming value, and a sequential minimal optimization value.

### Claim 16 (depends on 10)

The method of claim 10, wherein when the relevance value of the utterance is below the given threshold value, the utterance is considered a non-essential utterance with respect to completing the task and is removed from the dialog model.

### Claim 17 (depends on 16)

The method of claim 16, wherein when the relevance value of the utterance is not below the given threshold value, the utterance is considered an essential utterance with respect to completing the task and is included in the dialog model.

### Claim 18 (depends on 10)

The method of claim 10, wherein the relevance value is a measure of whether or not the utterance contains information relevant to completing the task.

### Claim 19

An article of manufacture comprising a non-transitory computer-readable storage medium having embodied therein executable program code that when executed by a processor causes the processor to perform steps of:
identifying one or more concepts from a document, wherein the document comprises information on completing a task;
computing a relevance value of an utterance with respect to completing the task using the one or more identified concepts; and
removing the utterance from a dialog model to be used for completing the task when the relevance value of the utterance is below a given threshold value.

### Claim 20 (depends on 19)

The article of claim 19, wherein the relevance value is a measure of whether or not the utterance contains information relevant to completing the task.
