Method and apparatus for transferring from robot customer service to human customer service

Patent metadata

Field	Value
Publication number	US-10977664-B2
Application number	US-202016888801-A
Country	US
Kind code	B2
Filing date	May 31, 2020
Priority date	Jan 26, 2018
Publication date	Apr 13, 2021
Grant date	Apr 13, 2021

Abstract

Methods, systems, and devices, including computer programs encoded on computer storage media for transferring a robot customer service to a human customer service are provided. One of the methods includes: obtaining conversation characteristics from at least one round of conversations between the robot customer service and a customer; obtaining state characteristics of the customer; inputting the conversation characteristics and the state characteristics into a confidence score evaluation model to obtain a confidence score evaluation value; and when the confidence score evaluation value meets a robot-to-human intervention condition, transferring the customer to the human customer service. The confidence score evaluation model is a machine learning model, comprising a linear sub-model input with the conversation characteristics and a deep neural network sub-model input with the state characteristics.

First claim

The invention claimed is:

1. A method for transferring a robot customer service to a human customer service, comprising:
obtaining conversation characteristics from at least one round of conversations between the robot customer service and a customer;
obtaining state characteristics of the customer;
inputting the conversation characteristics and the state characteristics into a confidence score evaluation model to obtain a confidence score evaluation value, wherein the confidence score evaluation model is a machine learning model, comprising a linear sub-model input with the conversation characteristics and a deep neural network sub-model input with the state characteristics, wherein the state characteristics comprises at least one of action record characteristics, service state characteristics, and identity information characteristics of the customer, and wherein the deep neural network sub-model comprises a first dense neural layer to process the service state characteristics, a second dense neural layer to process the identity information characteristics, and a long short-term memory (LSTM) neural layer to process the action record characteristics; and
when the confidence score evaluation value meets a robot-to-human intervention condition, transferring the customer to the human customer service.

2. The method according to claim 1 , wherein:
the action record characteristics comprise at least one of the following: access records and operation records of the customer within a period of time;
the service state characteristics comprise at least one of the following: a service opening state, an account authentication state, an account login state, and an abnormal account state of a customer account; and
the identity information characteristics comprise at least one of the following: gender, age, and residential area of the customer.

3. The method according to claim 1 , wherein the conversation characteristics are natural language processing (NLP) characteristics, comprising one or more of the following: relevance between the customer's question and the robot customer service's answer, a number of rounds of questions and answers, and types of answers.

4. The method according to claim 1 , wherein the robot-to-human intervention condition comprises: the confidence score evaluation value being greater or less than a confidence score threshold.

5. The method according to claim 1 , wherein the confidence score evaluation model is trained using a conversation sample marked with a robot-to-human intervention point between the robot customer service and the customer and a sample of the state characteristics of the customer.

6. A device for transferring a robot customer service to a human customer service, comprising one or more processors and a non-transitory computer-readable memory coupled to the one or more processors and configured with instructions executable by the one or more processors to perform operations comprising:
obtaining conversation characteristics from at least one round of conversations between the robot customer service and a customer;
obtaining state characteristics of the customer;
inputting the conversation characteristics and the state characteristics into a confidence score evaluation model to obtain a confidence score evaluation value, wherein the confidence score evaluation model is a machine learning model, comprising a linear sub-model input with the conversation characteristics and a deep neural network sub-model input with the state characteristics, wherein the state characteristics comprises at least one of action record characteristics, service state characteristics, and identity information characteristics of the customer, and wherein the deep neural network sub-model comprises a first dense neural layer to process the service state characteristics, a second dense neural layer to process the identity information characteristics, and a long short-term memory (LSTM) neural layer to process the action record characteristics; and
when the confidence score evaluation value meets a robot-to-human intervention condition, transferring the customer to the human customer service.

7. The device according to claim 6 , wherein:
the action record characteristics comprise at least one of the following: access records and operation records of the customer within a period of time;
the service state characteristics comprise at least one of the following: a service opening state, an account authentication state, an account login state, and an abnormal account state of a customer account; and
the identity information characteristics comprise at least one of the following: gender, age, and residential area of the customer.

8. The device according to claim 6 , wherein the conversation characteristics are natural language processing (NLP) characteristics, comprising one or more of the following: relevance between the customer's question and the robot customer service's answer, a number of rounds of questions and answers, and types of answers.

9. The device according to claim 6 , wherein the robot-to-human intervention condition comprises: the confidence score evaluation value being greater or less than a confidence score threshold.

10. The device according to claim 6 , wherein the confidence score evaluation model is trained using a conversation sample marked with a robot-to-human intervention point between the robot customer service and the customer and a sample of the state characteristics of the customer.

11. A non-transitory computer-readable storage medium for transferring a robot customer service to a human customer service, storing instructions executable by one or more processors to cause the one or more processors to perform operations comprising:
obtaining conversation characteristics from at least one round of conversations between the robot customer service and a customer;
obtaining state characteristics of the customer;
inputting the conversation characteristics and the state characteristics into a confidence score evaluation model to obtain a confidence score evaluation value, wherein the confidence score evaluation model is a machine learning model, comprising a linear sub-model input with the conversation characteristics and a deep neural network sub-model input with the state characteristics, wherein the state characteristics comprises at least one of action record characteristics, service state characteristics, and identity information characteristics of the customer, and wherein the deep neural network sub-model comprises a first dense neural layer to process the service state characteristics, a second dense neural layer to process the identity information characteristics, and a long short-term memory (LSTM) neural layer to process the action record characteristics; and
when the confidence score evaluation value meets a robot-to-human intervention condition, transferring the customer to the human customer service.

12. The non-transitory computer-readable storage medium according to claim 11 , wherein:
the action record characteristics comprise at least one of the following: access records and operation records of the customer within a period of time;
the service state characteristics comprise at least one of the following: a service opening state, an account authentication state, an account login state, and an abnormal account state of a customer account; and
the identity information characteristics comprise at least one of the following: gender, age, and residential area of the customer.

13. The non-transitory computer-readable storage medium according to claim 11 , wherein the conversation characteristics are natural lan

Assignees

• Advanced New Technologies Co Ltd

Inventors

• Yang Minghui

Classifications

• G06N5/01

  Dynamic search techniques; Heuristics; Dynamic trees; Branch-and-bound · CPC title

• G06N3/045

  Combinations of networks · CPC title

• G06N3/044

  Recurrent networks, e.g. Hopfield networks · CPC title

• G06N3/09

  Supervised learning · CPC title

• G06N3/0442

  characterised by memory or gating, e.g. long short-term memory [LSTM] or gated recurrent units [GRU] · CPC title