# US 10,872,697 B2 · Triage dialogue method, device and system

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

- Number: US 10,872,697 B2
- Document ID: US10872697B2
- Status: Granted
- Kind: B2
- Application number: 16/082,031
- Filed: 2018-01-10
- Published: 2019-12-19
- Granted: 2020-12-22
- Earliest claimed: 2017-06-28
- Assignees: Boe Technology Group Co., Ltd.
- Applicants: Boe Technology Group Co., Ltd.
- Inventors: Zhenzhong Zhang
- Web page: https://patentgrep.com/patent/US10872697B2

## CPC

- G16H 50/20: for computer-aided diagnosis, e.g. based on medical expert systems
- G16H 40/67: for remote operation
- H04L 67/10: in which an application is distributed across nodes in the network (software deployment G06F8/60; multiprogramming arrangements G06F9/46)
- G16H: HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- H04L: TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION (arrangements common to telegraphic and telephonic communication H04M)

## Abstract

A triage dialogue method, a device and a system are provided. The triage dialogue method includes: receiving triage dialogue content; dividing the triage dialogue content into a series of dialogue features, and generating dialogue feature vectors in one-to-one correspondence with the dialogue features; inputting the dialogue feature vectors into an encoder, and generating an intermediate representation vector through the encoder; determining whether the triage task is successfully completed according to the intermediate representation vector; and adjusting the triage dialogue policy according to the determination result of the triage task.

## Claims

### Claim 1

A triage dialogue method, comprising:
receiving triage dialogue content;
dividing the triage dialogue content into a series of dialogue features, and generating dialogue feature vectors in one-to-one correspondence with the dialogue features;
inputting the dialogue feature vectors into an encoder, and generating an intermediate representation vector through the encoder;
determining whether a triage task is successfully completed according to the intermediate representation vector; and
adjusting a triage dialogue policy according to a determination result of the triage task,
wherein adjusting the triage dialogue policy according to the determination result of the triage task includes:
continuing to interact with a user when the determination result is that the triage task is determined to have failed;
prompting the user to provide feedback information and determining whether to continue to interact with the user according to the feedback information when the determination result is that whether the triage task is successfully completed is uncertain; or
ending the dialogue with the user when the determination result is that the triage task is determined to have been completed,
wherein prompting the user to provide feedback information and determining whether to continue to interact with the user according to the feedback information when the determination result is that whether the triage task is successfully completed is uncertain includes:
ending the dialogue with the user when the feedback information indicates that the triage task has been completed; or
selecting information with the greatest information gain to interact with the user when the feedback information indicates that the triage task has not been completed.

### Claim 2 (depends on 1)

The triage dialogue method according to claim 1, wherein
the encoder adopts bidirectional gated recurrent unit (BGRU) network; and
based on a forward hidden sequence and a backward hidden sequence of the BGRU network, the intermediate representation vector h is:
h
=
1
T
⁢
∑
i
=
1
T
⁢
m
i
,
in which mi=[{right arrow over (m)}i;i] represents a concatenation between the ist element {right arrow over (m)}i in the forward hidden sequence {right arrow over (m)}1:T and the ist element in the backward hidden sequence T:1; and T represents the number of the dialogue features.

### Claim 3 (depends on 1)

The triage dialogue method according to claim 1, wherein determining whether the triage task is successfully completed according to the intermediate representation vector includes:
calculating success probability of the completion of the triage task based on previous intermediate representation vectors and the intermediate representation vector.

### Claim 4 (depends on 3)

The triage dialogue method according to claim 3, wherein
determining whether the triage task is successfully completed is characterized as Gaussian process; and the success probability is:
p
⁡
(
y
=
1
|
h
,
H
)
=
ϕ
(
μ
*
1
+
σ
*
2
)
,
in which H represents the previous intermediate representation vectors; h represents the intermediate representation vector and is the current intermediate representation; y∈{−1, 1}, and y=1 represents the successful completion of the triage task; φ(⋅) represents the distribution function of standard normal distribution; μ* and σ*2 respectively represent a posterior mean and a posterior variance of ƒ(h); and f(⋅) represents a potential function which is characterized as the Gaussian process.

### Claim 5 (depends on 3)

The triage dialogue method according to claim 3, wherein
when the success probability falls into a failure interval, the determination result is that the triage task is determined to have failed;
when the success probability falls into an uncertainty interval, the determination result is that whether the triage task is successfully completed is uncertain; or
when the success probability falls into a success interval, the determination result is that the triage task is determined to have been completed.

### Claim 6 (depends on 1)

The triage dialogue method according to claim 1, wherein selecting the information with the greatest information gain to interact with the user includes:
calculating information gains of different symptoms, and determining the symptom relevant to the greatest information gain; and
selecting the symptom relevant to the greatest information gain to interact with the user.

### Claim 7

A triage dialogue device, comprising a processor and a memory, wherein the memory is configured to store instructions; and the following operations are implemented when the instructions are executed by the processor:
receiving triage dialogue content;
dividing the triage dialogue content into a series of dialogue features, and generating dialogue feature vectors in one-to-one correspondence with the dialogue features;
inputting the dialogue feature vectors into an encoder, and generating an intermediate representation vector through the encoder;
determining whether a triage task is successfully completed according to the intermediate representation vector; and
adjusting a triage dialogue policy according to a determination result of the triage task,
wherein adjusting the triage dialogue policy according to the determination result of the triage task includes:
continuing to interact with a user when the determination result is that the triage task is determined to have failed;
prompting the user to provide feedback information and determining whether to continue to interact with the user according to the feedback information when the determination result is that whether the triage task is successfully completed is uncertain; or
ending the dialogue with the user when the determination result is that the triage task is determined to have been completed,
wherein prompting the user to provide feedback information and determining whether to continue to interact with the user according to the feedback information when the determination result is that whether the triage task is successfully completed is uncertain includes:
ending the dialogue with the user when the feedback information indicates that the triage task has been completed; or
selecting information with the greatest information gain to interact with the user when the feedback information indicates that the triage task has not been completed.

### Claim 8 (depends on 7)

The triage dialogue device according to claim 7, wherein
the encoder adopts BGRU network; and
based on a forward hidden sequence and a backward hidden sequence of the BGRU network, the intermediate representation vector h is:
h
=
1
T
⁢
∑
i
=
1
T
⁢
m
i
,
in which mi=[{right arrow over (m)}i;i] represents a concatenation between the ist element {right arrow over (m)}i in the forward hidden sequence {right arrow over (m)}1:T and the ist element in the backward hidden sequence T:1; and T represents the number of the dialogue features.

### Claim 9 (depends on 7)

The triage dialogue device according to claim 7, wherein determining whether the triage task is successfully completed according to the intermediate representation vector includes:
calculating success probability of the completion of the triage task based on the previous intermediate representation vectors and the intermediate representation vector.

### Claim 10 (depends on 9)

The triage dialogue device according to claim 9, wherein determining whether the triage task is successfully completed is characterized as Gaussian process; and the success probability is:
p
⁡
(
y
=
1
|
h
,
H
)
=
ϕ
(
μ
*
1
+
σ
*
2
)
,
in which H represents the previous intermediate representation vectors; h represents the intermediate representation vector and is the current intermediate representation; y∈{−1, 1}, and y=1 represents the successful completion of the triage task; φ(⋅) represents the distribution function of standard normal distribution; μ* and σ*2 respectively represent the posterior mean and the posterior variance of ƒ(h); and f(⋅) represents the potential function which is characterized as the Gaussian process.

### Claim 11 (depends on 9)

The triage dialogue device according to claim 9, wherein
when the success probability falls into a failure interval, the determination result is that the triage task is determined to have failed;
when the success probability falls into an uncertainty interval, the determination result is that whether the triage task is successfully completed is uncertain; or
when the success probability falls into a success interval, the determination result is that the triage task is determined to have been completed.

### Claim 12 (depends on 7)

The triage dialogue device according to claim 7, wherein selecting the information with the greatest information gain to interact with the user includes:
calculating information gains of different symptoms, and determining the symptom relevant to the greatest information gain; and
selecting the symptom relevant to the greatest information gain to interact with the user.

### Claim 13 (depends on 7)

The triage dialogue device according to claim 7, further comprising an interactive interface, wherein the interactive interface is configured to interact with the user.

### Claim 14

A non-transitory computer-readable storage medium, wherein computer instructions are stored on the computer readable storage medium; and the following operations are implemented when the computer instructions are executed by a processor:
receiving triage dialogue content;
dividing the triage dialogue content into a series of dialogue features, and generating dialogue feature vectors in one-to-one correspondence with the dialogue features;
inputting the dialogue feature vectors into an encoder, and generating an intermediate representation vector through the encoder;
determining whether a triage task is successfully completed according to the intermediate representation vector; and
adjusting a triage dialogue policy according to a determination result of the triage task,
wherein adjusting the triage dialogue policy according to the determination result of the triage task includes:
continuing to interact with a user when the determination result is that the triage task is determined to have failed;
prompting the user to provide feedback information and determining whether to continue to interact with the user according to the feedback information when the determination result is that whether the triage task is successfully completed is uncertain; or
ending the dialogue with the user when the determination result is that the triage task is determined to have been completed,
wherein prompting the user to provide feedback information and determining whether to continue to interact with the user according to the feedback information when the determination result is that whether the triage task is successfully completed is uncertain includes:
ending the dialogue with the user when the feedback information indicates that the triage task has been completed; or
selecting information with the greatest information gain to interact with the user when the feedback information indicates that the triage task has not been completed.