Human-computer interaction method, device, computer equipment and storage medium

By detecting and responding to the voice information of the target user, the intelligent robot pauses and continues to play in the noise environment, solving the problem of interrupting the intelligent robot in the noise environment, improving the flexibility and intelligence of human-computer interaction, and improving interaction efficiency.

CN111752523BActive Publication Date: 2024-11-26SHENZHEN ZHUIYI TECH CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202010400842.1
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2020-05-13
Publication Date
2024-11-26
Estimated Expiration
2040-05-13

AI Technical Summary

Technical Problem

Existing intelligent robots have difficulty responding to user interruptions in noisy environments, resulting in speeches being interrupted by mistake or replaying the speech from scratch, and human-computer interaction is not flexible and intelligent enough.

Method used

By detecting the voice information in the target user environment, pausing and recording the playback position, identifying the semantics of the voice information and performing corresponding operations, continuing to play from the paused position after the response is completed, and using the preset sound feature detection model and deep learning algorithm to identify the user's voice .

Benefits of technology

It improves the flexibility and intelligence of human-computer interaction, avoids the incomplete voice broadcast and information transmission caused by user interruption, and improves interaction efficiency.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN111752523B_ABST
    Figure CN111752523B_ABST
Patent Text Reader

Abstract

The present application relates to a human-computer interaction method, device, computer equipment and storage medium. When playing the target voice data, the telephone robot detects the voice information in the target user's environment. If the voice information is the target user's voice information, the target voice data is paused, and the playback position of the target voice data at the pause time is recorded. Then, the semantics of the target user's voice information is identified, and a response operation corresponding to the semantics is performed. After the response operation is completed, the remaining content of the target voice data is played from the playback position at the pause time. This method can avoid the inefficiency and incomplete information transmission caused by repeated voice broadcasting due to user interruptions, improve the robot's flexibility in responding to user interruptions, and make human-computer interaction more efficient, flexible and intelligent.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present application relates to the field of artificial intelligence technology, and in particular to a human interaction method, apparatus, computer device and storage medium. Background Art

[0002] With the development of artificial intelligence technology, intelligent robots have been applied in various fields. For example, they provide automatic customer service, intelligent marketing, content navigation, intelligent voice control, entertainment chat and other types of services in many fields such as telecommunications operators, financial services, e-government, e-commerce, various smart terminals and personal Internet information services.

[0003] Most intelligent robots can have basic communication with users during the service process. However, when the robot is playing the script, if the environment is very noisy, such as other people talking around the user, the robot's speech will be interrupted by mistake; or the robot will mechanically replay the script from the beginning after being interrupted by the user multiple times.

[0004] Therefore, existing robots have the problem of insufficient flexibility and intelligence in human-computer interaction. Summary of the invention

[0005] Based on this, it is necessary to provide a human-computer interaction method, device, computer equipment and storage medium that can make human-computer interaction more flexible and intelligent in response to the above technical problems.

[0006] In a first aspect, a human-computer interaction method is provided, the method comprising:

[0007] Detecting voice information in the target user's environment when playing the target voice data;

[0008] If the voice information is the voice information of the target user, pausing the playing of the target voice data and recording the playing position of the target voice data at the pausing time;

[0009] Identify the semantics of the target user's voice information and perform a response operation corresponding to the semantics;

[0010] After the response operation is completed, the remaining content of the target voice data is played from the playback position at the pause time.

[0011] In one embodiment, the above-mentioned recognition of the semantics of the target user's voice information and execution of a response operation corresponding to the semantics include:

[0012] Determine the information type of the voice information according to the semantics of the voice information of the target user;

[0013] According to the information type, a response operation corresponding to the semantics is performed.

[0014] In one embodiment, the above-mentioned execution of the response operation corresponding to the semantics according to the information type includes: if the information type is an inquiry type, matching the reply content corresponding to the semantics of the voice information of the target user from a preset knowledge base, and playing the reply content; the knowledge base includes reply content corresponding to multiple information types; or,

[0015] If the information type is an echo type, after pausing for a preset time, continue to play the remaining content of the target voice data from the playback position at the pause time.

[0016] In one embodiment, the playback position of the target voice data at the pause time is recorded, including:

[0017] Determine the sentence corresponding to the pause moment in the target speech data;

[0018] The position of the sentence corresponding to the pause moment in the target voice data is determined as the playback position of the target voice data at the pause moment.

[0019] In one embodiment, after detecting the voice information in the target user's environment when playing the target voice data, the method includes:

[0020] If the voice information is not the target user's voice information, no response is made to the voice information, and the target voice data continues to be played.

[0021] In one embodiment, the method further comprises:

[0022] The preset sound feature detection model is used to identify whether the voice information in the target user's environment is the voice information of the target user.

[0023] In one embodiment, the training process of the sound feature detection model includes:

[0024] Acquire multiple sample voice features and user identifiers corresponding to each sample voice feature;

[0025] Through a deep learning algorithm, the correspondence between each voice feature and the corresponding user identifier is learned to iteratively train the initial sound feature detection model until the change amplitude of the value of the loss function of the sound feature detection model is less than a preset threshold, thereby obtaining a sound feature detection model.

[0026] In a second aspect, the present application provides a human-computer interaction device, the device comprising:

[0027] A detection module, used to detect voice information in the target user's environment when playing the target voice data;

[0028] A stop module, used to pause the playing of the target voice data if the voice information is the voice information of the target user, and record the playing position of the target voice data at the pause time;

[0029] A response module, used to recognize the semantics of the target user's voice information and perform a response operation corresponding to the semantics;

[0030] The playing module is used to play the remaining content of the target voice data from the playing position at the pause time after the response operation is completed.

[0031] In one embodiment, the response module includes:

[0032] A type determination unit, used to determine the information type of the voice information according to the semantics of the voice information of the target user;

[0033] The response execution unit is used to execute a response operation corresponding to the semantics according to the information type.

[0034] In one embodiment, the above-mentioned execution response unit is specifically used to match reply content corresponding to the semantics of the target user's voice information from a preset knowledge base and play the reply content if the information type is an inquiry type; the knowledge base includes reply content corresponding to multiple information types; or, if the information type is an echo type, after pausing for a preset period of time, continue to play the remaining content of the target voice data from the playback position at the pause time.

[0035] In one embodiment, the stop module comprises:

[0036] A sentence determination unit, used to determine the sentence corresponding to the pause moment in the target speech data;

[0037] The position determination unit is used to determine the position of the sentence corresponding to the pause moment in the target voice data as the playback position of the target voice data at the pause moment.

[0038] In one embodiment, the above-mentioned response module is further specifically used for not responding to the voice information if the voice information is voice information of a non-target user, and continuing to play the target voice data.

[0039] In one embodiment, the detection module is specifically used to identify whether the voice information in the target user's environment is the target user's voice information through a preset sound feature detection model.

[0040] In one embodiment, the apparatus further comprises:

[0041] The sample acquisition module is used to obtain a variety of sample voice features and the user ID corresponding to each sample voice feature.

[0042] The model training module is used to learn the correspondence between each voice feature and the corresponding user identifier through a deep learning algorithm, so as to iteratively train the initial sound feature detection model until the change amplitude of the value of the loss function of the sound feature detection model is less than a preset threshold, thereby obtaining a sound feature detection model.

[0043] In a third aspect, the present application provides a computer device, comprising a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the steps of the human-computer interaction method of any one of the embodiments of the first aspect are implemented.

[0044] In a fourth aspect, the present application provides a computer-readable storage medium having a computer program stored thereon, and when the computer program is executed by a processor, the steps of the human-computer interaction method of any one of the embodiments of the first aspect above are implemented.

[0045] The human-computer interaction method, device, computer equipment and storage medium provided by the present application are as follows: when the telephone robot is playing the target voice data, it detects the voice information in the environment where the target user is located. If the voice information is the voice information of the target user, it pauses the playing of the target voice data, and records the playing position of the target voice data at the pause time, then identifies the semantics of the voice information of the target user, performs the response operation corresponding to the semantics, and after the response operation is completed, plays the remaining content of the target voice data from the playing position at the pause time. In this method, after the telephone robot detects the voice information of the target user when playing the target voice data, it first pauses the playing of the target voice data, and makes a corresponding response to the target user according to the semantics of the voice information output by the current target user, and then continues to play the remaining target voice data from the pause position. That is to say, after being interrupted by the target user, the telephone robot first responds according to the interruption intention of the target user. After the response is completed, it does not need to play the target voice data from the beginning, and only plays the remaining content. This avoids the inefficiency and incomplete information transmission caused by repeated voice broadcasting due to the user's interruption, improves the robot's flexible response to the user's interruption, and makes the human-computer interaction more efficient, more flexible and intelligent. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] Figure 1 An application environment diagram of a human-computer interaction method provided by an embodiment;

[0047] Figure 1a A diagram of the internal structure of a telephone robot provided for one embodiment;

[0048] Figure 2 A schematic diagram of a flow chart of a human-computer interaction method provided by an embodiment;

[0049] Figure 3A flowchart of a human-computer interaction method provided by another embodiment;

[0050] Figure 4 A flowchart of a human-computer interaction method provided by another embodiment;

[0051] Figure 5 A schematic diagram of a human-computer interaction method provided by an embodiment;

[0052] Figure 6 A structural block diagram of a human-computer interaction device provided by an embodiment;

[0053] Figure 7 A structural block diagram of a human-computer interaction device provided by another embodiment;

[0054] Figure 8 FIG. 4 is a diagram showing the internal structure of a computer device in one embodiment. DETAILED DESCRIPTION

[0055] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0056] See also Figure 1 The present application provides an application environment of a human-computer interaction method, wherein a telephone robot 01 can perform voice interaction with a user, and the interaction scenarios include but are not limited to online question-and-answer, consultation, command execution and other service scenarios. The telephone robot 01 includes but is not limited to outbound call robots, chat robots, intelligent customer service, intelligent assistants and other service-type robots. The internal structure of the telephone robot can be found in FIG. Figure 1a As shown, the telephone robot includes a processor, a memory, a network interface and a database connected through a system bus. The processor is used to provide computing and control capabilities. The memory includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database is used to store data of a human-computer interaction method. The network interface is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, a human-computer interaction method is implemented. It can be understood that Figure 1a The internal structure of the intelligent robot shown is only an example and is not intended to be limiting.

[0057] The embodiments of the present application provide a human-computer interaction method, device, computer equipment and storage medium, making human-computer interaction more flexible and intelligent. The technical solution of the present application and how the technical solution of the present application solves the above-mentioned technical problems will be described in detail below through embodiments and in combination with the accompanying drawings. The following specific embodiments can be combined with each other, and the same or similar concepts or processes may not be repeated in some embodiments. It should be noted that the human-computer interaction method provided by the present application, Figure 2-Figure 5 The execution subject is the telephone robot. Figure 2-Figure 5 The execution subject may also be a human-computer interaction device, which may be implemented as part or all of the telephone robot through software, hardware, or a combination of software and hardware.

[0058] In order to make the purpose, technical solutions and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be clearly and completely described below in combination with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments.

[0059] In one embodiment, Figure 2 A human-computer interaction method is provided. This embodiment relates to a specific process in which a telephone robot, when playing target voice data, pauses the playback if it detects the voice information of a target user, and performs a corresponding operation according to the semantics of the voice information, and after the operation is completed, continues to play the target voice data from the position where the playback was paused, such as Figure 2 As shown, the method includes:

[0060] S101, detecting voice information in the environment where the target user is located when playing target voice data.

[0061] Among them, the target voice data refers to the voice data that the telephone robot is playing at the current time, for example, the telephone robot is playing important information notifications, or playing marketing information in a marketing scenario.

[0062] Since the telephone robot is an intelligent voice service device, it can accurately understand the user's intentions and questions through artificial intelligence technologies such as voice recognition, voice synthesis, semantic understanding, and natural language generation, and provide autonomous online question-and-answer, consultation, and command execution services through natural and smooth human-computer interaction. Therefore, in actual applications, when the telephone robot plays the target voice data, it is in a scene where it is interacting with the target user. In this scene, the telephone robot can detect the voice information in the target user's environment through voice detection, for example, whether these voice information belong to the voice output of the target user, whether these voice information are voice information output by the target user to the telephone robot or voice information output to other objects, etc.

[0063] Among them, the telephone robot needs to detect and collect voice information in the target user's environment. The main purpose of collecting voice information in the target user's environment is to detect whether there is voice information of the target user in the collected voice information. It should be understood that in actual applications, the voice information in the target user's environment includes not only the target user's voice information, but also the voice information of other people or objects. Therefore, the telephone robot only needs to detect whether there is information of the target user in the collected voice information.

[0064] Optionally, in one embodiment, the telephone robot can identify whether the voice information in the target user's environment is the voice information of the target user through a preset voice feature detection model. In one embodiment, the training process of the voice feature detection model includes: obtaining a plurality of sample voice features and a user identifier corresponding to each sample voice feature; learning the correspondence between each voice feature and the corresponding user identifier through a deep learning algorithm to iteratively train the initial voice feature detection model until the change amplitude of the value of the loss function of the voice feature detection model is less than a preset threshold, thereby obtaining a voice feature detection model.

[0065] The sound feature detection model is a pre-trained model used to detect the features of different sounds in a voice message, for example, the features of the sound are timbre, voice pattern, etc. Therefore, the telephone robot can input the collected voice information of the target user's environment into the sound feature detection model, and judge whether the collected voice information contains the target user's voice information from the output result of the model. When training, the sound feature detection model uses a variety of sample voice features and the user identifiers corresponding to each sample voice feature as a training set for training. Of course, in order to ensure that the sound feature detection model has a higher accuracy in recognizing the sound, when selecting the training set, the samples are diversified and rich as much as possible, so that the sound feature detection model can learn more comprehensive sound features. During training, the value of the loss function is used as a guide for the direction of model training. When the loss function value tends to be stable, that is, the change range of the loss function value for multiple consecutive times is very small, for example, the change range is less than 0.1, it is determined that the training of the sound feature detection model is completed, and a trained sound feature detection model is obtained. In this way, the voice information of the target user can be quickly and accurately detected by the pre-trained sound feature detection model.

[0066] Optionally, the telephone robot can identify whether the voice information in the target user's environment is the target user's voice information by comparing it with pre-stored voice characteristics of the user. That is, when the telephone robot interacts with the user, it has already stored the voice characteristics of the target user in advance. In this way, when detecting the target user's voice information, the various sound characteristics of the voice information in the environment can be compared with the voice characteristics of the target user. By comparison, it can be determined whether there is voice information of the target user in the environment.

[0067] S102: If the voice information is the voice information of the target user, pause the playing of the target voice data, and record the playing position of the target voice data at the pause time.

[0068] The scenario in the actual application corresponding to this step is that when the telephone robot is playing a piece of voice data, it is suddenly interrupted by the voice information input by the target user. That is, the telephone robot detects that the voice information in the environment where the above-mentioned target user is located is the voice information of the target user, or the voice information in the environment where the above-mentioned target user is located includes the voice information of the target user.

[0069] Then, in the case of being suddenly interrupted by the voice message input by the target user, the telephone robot pauses the target voice data currently being played. When the telephone robot pauses, it records the playback position of the target voice data at the time of the pause. For example, the target voice data played by the telephone robot is a piece of insurance information. When it is playing "This insurance is insured for 10 years and only requires 5 years of payment, pay every month", the telephone robot detects the voice message of the target user. At this time, the playback position of the target voice data at the time of the pause is the position between "This insurance is insured for 10 years and only requires 5 years of payment" and "Pay every month".

[0070] S103, identifying the semantics of the target user's voice information, and executing a response operation corresponding to the semantics.

[0071] After the telephone robot pauses the target voice data being played and records the pause position, the telephone robot needs to identify the semantics of the target user's voice information, that is, it needs to identify the target user's true intention of inputting voice information in order to perform corresponding response operations according to the target user's true intention.

[0072] It should be noted here that the process of the telephone robot recognizing the semantics of the target user's voice information and the process of the telephone robot pausing the target voice data being played and recording the pause position can be carried out simultaneously or sequentially. For example, in actual applications, these two processes can be carried out simultaneously, or they can be carried out sequentially in a way that can quickly complete semantic recognition and record the pause position. In this way, the telephone robot's response process can be ensured to be smooth, avoiding users from sensing the pause time of the telephone robot, and improving the human-computer interaction experience.

[0073] For example, the telephone robot can use a pre-trained semantic recognition network model to recognize the semantics of the target user's voice information. Of course, the semantic recognition network model can be two different models independent of the above-mentioned sound feature detection model, and different models are responsible for different detection functions; but it can also be the same model, that is, the above-mentioned sound feature detection model can not only detect the sound features, but also detect the specific semantic content of the target user's voice information. This embodiment does not make specific limitations on this, and different network models can be adopted according to actual conditions.

[0074] After recognizing the semantics of the target user's voice information, the telephone robot needs to execute a response operation corresponding to the semantics. For example, the telephone robot can match the corresponding response reply from a preset knowledge base according to the semantics of the target user's voice information and execute the response reply. Other methods are also possible, and this embodiment does not limit this.

[0075] S104, after the response operation is completed, the remaining content of the target voice data is played from the playback position at the pause time.

[0076] After the response operation is completed, the telephone robot re-acquires the playback position of the target voice data just recorded at the pause time, and then continues to play the remaining content of the target voice data from the playback position. For example: the playback position of the target voice data recorded by the telephone robot at the pause time is the position between "This insurance is insured for 10 years and only requires 5 years of payment" and "Pay every month". At this time, the telephone robot continues to play from this position, that is, "Pay 200 yuan per month. Can I get you one?" In this way, the remaining content continues to be played from the position where it was paused, without having to start from the beginning, thereby improving the efficiency of human-computer interaction.

[0077] The human-computer interaction method provided in this embodiment is that the telephone robot detects the voice information in the environment where the target user is located when playing the target voice data. If the voice information is the voice information of the target user, the target voice data is paused, and the playback position of the target voice data at the pause time is recorded, and then the semantics of the target user's voice information are identified, and the response operation corresponding to the semantics is performed. After the response operation is completed, the remaining content of the target voice data is played from the playback position at the pause time. In this method, since the telephone robot detects the voice information of the target user when playing the target voice data, the target voice data is first paused, and a corresponding response is made to the target user according to the semantics of the voice information output by the current target user, and then the remaining target voice data will continue to be played from the pause position. That is to say, after being interrupted by the target user, the telephone robot first responds according to the interruption intention of the target user. After the response is completed, there is no need to play the target voice data from the beginning, and only the remaining content is played. This avoids the inefficiency and incomplete information transmission caused by repeated voice broadcasting due to the user's interruption, improves the robot's flexible response to the user's interruption, and makes the human-computer interaction more efficient, flexible and intelligent.

[0078] In the above embodiment, the voice information in the environment where the target user is located is described as the target voice. However, in actual applications, the voice information in the environment must also contain voice information of non-target users. The following is an example of an example in which the voice information in the environment is the voice information of non-target users. In one embodiment, if the voice information is the voice information of a non-target user, no response is made to the voice information, and the target voice data continues to be played.

[0079] In this embodiment, if the voice information in the target user's environment detected by the telephone robot is the voice information of a non-target user, for example, the telephone robot detects through the sound feature detection model in the above embodiment that the voice information in the target user's environment is the voice information of a person other than the target user, or is the voice information emitted by other objects in the environment, then in this case, the telephone robot can ignore the voice information, do not make any response, and continue to play the target voice data. In this way, the telephone robot is avoided from being interrupted by noise or other sounds in the environment, thereby improving the intelligence of the telephone robot in human-computer interaction.

[0080] Regarding the process of the telephone robot recognizing the semantics of the target user's voice information and executing the response operation corresponding to the semantics, this application provides a specific embodiment for detailed description. In one embodiment, Figure 3 As shown, the above step S103 includes:

[0081] S201, determining the information type of the voice information according to the semantics of the voice information of the target user.

[0082] The information type of the voice information is a classification of the type of intention expressed by the voice information of the target user. For example, the information type may be an inquiry type, an acknowledgment type, an argument type, a praise type, and the like.

[0083] In actual applications, the telephone robot can determine the actual intention of the voice information based on the semantics of the voice information, and then classify the information type according to the intention. For example, the correspondence between multiple semantics and information types can be pre-set. After determining the semantics of the target user's voice information, the telephone robot matches the determined semantics with the pre-set correspondence, and determines the information type corresponding to the semantics based on the match. This embodiment does not specifically limit this method.

[0084] S202: Execute a response operation corresponding to the semantics according to the information type.

[0085] After determining the specific information type, the telephone robot performs the corresponding response operation. Here, different information types correspond to different response operations, so that the telephone robot can flexibly respond to the target user and improve the intelligence of human-computer interaction.

[0086] The following provides two specific response operations corresponding to information types. Optionally, in one embodiment, the above-mentioned response operation corresponding to the semantics according to the information type includes: if the information type is an inquiry type, matching the reply content corresponding to the semantics of the target user's voice information from a preset knowledge base, and playing the reply content; the knowledge base includes reply content corresponding to multiple information types.

[0087] The information type of inquiry type indicates that the current target user has doubts about the voice content played by the telephone robot, and therefore asks questions. After the telephone robot determines the specific question based on the semantics of the target user's voice information, it matches the reply content corresponding to the semantics from the pre-established knowledge base and plays the reply content. The knowledge base here is pre-established and stores a large amount of reply content corresponding to the semantics of different voice information in the information type.

[0088] For example, a telephone robot is playing the target voice data "This insurance covers 10 years and only requires payment for 5 years, with monthly payments", and suddenly detects that the target user has sent a voice message, and the voice message is specifically "What is the name of the insurance you just mentioned?" Then, after determining the specific semantics of the voice message, the telephone robot matches the corresponding reply from the knowledge base as "The name of this insurance is Lifetime Health Insurance."

[0089] In this embodiment, when the telephone robot is playing a voice content and receives an interrogative voice from a target user, the telephone robot will promptly match a reply corresponding to the target user's intention in the knowledge base and respond to the target user based on the matched reply, so that the user can quickly get the answer he wants, thereby improving the efficiency of human-computer interaction.

[0090] Optionally, in another embodiment, the above-mentioned response operation corresponding to the semantics is performed according to the information type, including: if the information type is an echo type, after pausing for a preset time, continue to play the remaining content of the target voice data from the playback position at the pause time.

[0091] In this embodiment, the information type is an acknowledgment type, which means that the voice information sent by the target user is just a simple acknowledgment, for example, um, oh, okay, like this, yes, I understand, etc.

[0092] If the telephone robot detects the target user's voice information and recognizes that the voice information is just a simple echoing voice, the telephone robot does not need to answer anything. It only needs to pause for a preset time and then continue to play the remaining content of the target voice data from the playback position at the pause time, for example, continue to play the voice content being played after 2 seconds. For example, the telephone robot is playing the target voice data "This insurance is insured for 10 years and only requires 5 years of payment. The monthly payment amount is", and suddenly detects that the target user has sent a voice message, and the voice message is specifically "Oh, oh, so", then the telephone robot will pause for 2 seconds and continue to play the subsequent content from the pause position.

[0093] Of course, in actual applications, the telephone robot can also reply with some simple replies. For example, if the target user sends a voice message of "Oh, oh, so", the telephone robot replies "Well". This embodiment does not limit this. As long as the telephone robot simply responds after detecting the target user's affirmative information, there is no need to make too many reply operations, so as to avoid wasting time or causing incomplete playback of the target voice data.

[0094] Regarding the process of the telephone robot recording the pause position of the target voice data, an embodiment is provided for specific description, such as Figure 4 As shown, in one embodiment, the above step S102 includes:

[0095] S301, determining a sentence corresponding to a pause moment in target speech data.

[0096] The target voice data is a piece of voice content that the telephone robot is playing, and a piece of voice content is composed of many sentences. Therefore, after the telephone robot pauses playing the target voice data, it determines the sentence to be played at this pause time. For example, the telephone robot plays "This insurance is insured for 10 years and only requires 5 years of payment, pay every month..." Then the sentence at the pause time is "Pay every month."

[0097] S302: Determine the position of the sentence corresponding to the pause moment in the target voice data as the playback position of the target voice data at the pause moment.

[0098] After determining the sentence corresponding to the pause moment in the target voice data, the telephone robot determines the position of the sentence in the target voice data as the playback position of the target voice data at the pause moment.

[0099] For example, if the sentence at the pause time is "pay every month", then the position of the sentence in the target voice data is after "this insurance is insured for 10 years and only requires 5 years of payment", which is the playback position of the target voice data at the pause time.

[0100] This embodiment determines the position of the sentence at the pause moment in the target voice data, and determines the position of the sentence in the target voice data as the playback position of the target voice data at the pause moment, and uses the sentence as the basis for determining the position, thereby improving the accuracy of recording the pause position.

[0101] In addition, the present application also provides a human-computer interaction method, such as Figure 5 As shown, this embodiment includes:

[0102] S1, plays the target voice data and simultaneously detects the voice information in the target user's environment;

[0103] S2, detect whether the voice information is the voice information of the target user; if so, execute S3, if not, execute S4;

[0104] S3, pausing the playing of the target voice data, and recording the playing position of the target voice data at the pause time;

[0105] S4, do not respond to the voice information, and continue to play the target voice data;

[0106] S5, recognizes the semantics of the target user’s voice information;

[0107] S6, determining the information type of the voice information according to the semantics of the voice information of the target user; and executing S7 and S8 respectively according to the information type;

[0108] S7, if the information type is an echo type, pausing for a preset time and then continuing to play the remaining content of the target voice data from the playback position at the pause time;

[0109] S8, if the information type is an inquiry type, matching the reply content corresponding to the semantics of the target user's voice information from a preset knowledge base, and playing the reply content; the knowledge base includes reply content corresponding to multiple information types;

[0110] S9, after the response operation is completed, the remaining content of the target voice data is played from the playback position at the pause time.

[0111] The implementation principles and technical effects of each step in all the human-computer interaction methods provided in the above embodiments are similar to those in the previous human-computer interaction method embodiments, and will not be repeated here.

[0112] It should be understood that although Figure 2-5 The steps in the flowchart are shown in sequence as indicated by the arrows, but these steps are not necessarily executed in the order indicated by the arrows. Unless otherwise specified in this document, there is no strict order restriction for the execution of these steps, and these steps can be executed in other orders. Moreover, Figure 2-5 At least part of the steps may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily executed at the same time, but can be executed at different times. The execution order of these sub-steps or stages is not necessarily sequential, but can be executed in turn or alternately with other steps or at least part of the sub-steps or stages of other steps.

[0113] In one embodiment, Figure 6 As shown, a human-computer interaction device is provided, including: a detection module 10, a stop module 11, a response module 12 and a play module 13, wherein:

[0114] A detection module 10, used to detect voice information in the environment where the target user is located when playing the target voice data;

[0115] The stop module 11 is used to pause the playing of the target voice data if the voice information is the voice information of the target user, and record the playing position of the target voice data at the pause time;

[0116] A response module 12, used to recognize the semantics of the target user's voice information and perform a response operation corresponding to the semantics;

[0117] The playing module 13 is used to play the remaining content of the target voice data from the playing position at the pause time after the response operation is completed.

[0118] In this embodiment, when the telephone robot is playing a section of target voice data, if the detection module detects that there is voice information in the environment where the target user is located, and the voice information is the voice of the target user, the stop module pauses the target voice data being played, and records the playback position of the target voice data at the pause time. After the stop module pauses the target voice data being played, the response module identifies the semantics of the voice information of the target user, and performs the corresponding response operation according to the parsed semantics. Regardless of which response operation is performed, after the response operation is completed, the playback module plays the remaining content of the target voice data from the playback position at the pause time. In this way, after being interrupted by the target user, the telephone robot first responds according to the interruption intention of the target user. After the response is completed, it does not need to play the target voice data from the beginning, and only plays the remaining content, avoiding the inefficiency and incomplete information transmission caused by repeated voice broadcasting due to the user's interruption, improving the robot's flexible response to the user's interruption, and making the human-computer interaction more efficient, more flexible and intelligent.

[0119] In one embodiment, Figure 7 As shown, the above-mentioned response module 12 includes: a type determination unit 121 and an execution response unit 122;

[0120] A type determination unit 121, configured to determine the information type of the voice information according to the semantics of the voice information of the target user;

[0121] The response execution unit 122 is used to execute a response operation corresponding to the semantics according to the information type.

[0122] In this embodiment, the type determination unit in the response module determines the actual intention of the voice information according to the semantics of the voice information, and then divides the information type according to the intention. For example, the information type can be an inquiry type, an acquiescent type, an argumentative type, a praising type, etc.; after determining the specific information type, the telephone robot performs the corresponding response operation, wherein different information types correspond to different response operations, so that the telephone robot can flexibly respond to the target user and improve the intelligence of human-computer interaction.

[0123] In one embodiment, the above-mentioned execution response unit is specifically used to match reply content corresponding to the semantics of the target user's voice information from a preset knowledge base and play the reply content if the information type is an inquiry type; the knowledge base includes reply content corresponding to multiple information types; or, if the information type is an echo type, after pausing for a preset period of time, continue to play the remaining content of the target voice data from the playback position at the pause time.

[0124] In this embodiment, two response operations corresponding to specific information types are provided. The information type is an inquiry type, which indicates that the current target user has questions about the voice content played by the telephone robot. The telephone robot needs to answer the questions, that is, match the reply content corresponding to the semantics of the target user's voice information from the preset knowledge base, and play the reply content, so that the user can quickly get the desired answer, thereby improving the efficiency of human-computer interaction; if the voice information sent by the target user is just a simple echo, for example, um, oh, okay, like this, yes, I know, etc., the telephone robot does not need to answer anything, but only needs to pause for a preset length of time and then continue to play the remaining content of the target voice data from the playback position at the pause time. It is a simple response without the need to make too many reply operations, thereby avoiding wasting time or causing incomplete playback of the target voice data.

[0125] In one embodiment, the stopping module 11 includes:

[0126] A sentence determination unit, used to determine the sentence corresponding to the pause moment in the target speech data;

[0127] The position determination unit is used to determine the position of the sentence corresponding to the pause moment in the target voice data as the playback position of the target voice data at the pause moment.

[0128] In this embodiment, after the telephone robot pauses playing the target voice data, it determines the sentence played at the pause time. For example, if the telephone robot plays "This insurance is insured for 10 years and only requires 5 years of payment, and the payment is made every month...", then the sentence at the pause time is "payment is made every month"; then the position of the sentence in the target voice data is determined as the playback position of the target voice data at the pause time. In this way, the accuracy of recording the pause position is improved by using the sentence as the basis for determining the position.

[0129] In one embodiment, the response module 12 is further configured to, if the voice information is not from a target user, not respond to the voice information and continue to play the target voice data.

[0130] In this embodiment, if the voice information is from a non-target user, the telephone robot can ignore the voice information, make no response, and continue to play the target voice data. This avoids interruptions to the telephone robot due to noise or other sounds in the environment, and improves the intelligence of the telephone robot in human-computer interaction.

[0131] In one embodiment, the detection module 10 is specifically used to identify whether the voice information in the target user's environment is the target user's voice information through a preset sound feature detection model.

[0132] In this embodiment, the sound feature detection model is pre-trained and is a model used to detect the features of different sounds in a piece of voice information. For example, the features of the sound are timbre, voice pattern, etc. The pre-trained sound feature detection model can be used to quickly and accurately detect the voice information of the target user.

[0133] In one embodiment, the apparatus further comprises:

[0134] The sample acquisition module is used to obtain a variety of sample voice features and the user ID corresponding to each sample voice feature.

[0135] The model training module is used to learn the correspondence between each voice feature and the corresponding user identifier through a deep learning algorithm, so as to iteratively train the initial sound feature detection model until the change amplitude of the value of the loss function of the sound feature detection model is less than a preset threshold, thereby obtaining a sound feature detection model.

[0136] In this embodiment, when training the sound feature detection model, multiple sample speech features and the user identifier corresponding to each sample speech feature are used as the training set for training. During training, the value of the loss function is used as the direction of model training to guide iterative training, which can make the training of the sound feature detection model more robust and accurate.

[0137] The implementation principles and technical effects of all the human-computer interaction devices provided in the above embodiments are similar to those of the above human-computer interaction method embodiments, and will not be repeated here.

[0138] For the specific definition of the human-computer interaction device, please refer to the definition of the human-computer interaction method above, which will not be repeated here. Each module in the above human-computer interaction device can be implemented in whole or in part by software, hardware and a combination thereof. The above modules can be embedded in or independent of the processor in the computer device in the form of hardware, or can be stored in the memory of the computer device in the form of software, so that the processor can call and execute the operations corresponding to the above modules.

[0139] In one embodiment, a computer device is provided. The computer device may be a terminal, and its internal structure diagram may be as described above. Figure 8As shown. The computer device includes a processor, a memory, a network interface, a display screen and an input device connected through a system bus. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The network interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, a human-computer interaction method is implemented. The display screen of the computer device can be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer device can be a touch layer covering the display screen, or a key, trackball or touchpad set on the computer device housing, or an external keyboard, touchpad or mouse, etc.

[0140] Those skilled in the art will appreciate that the above Figure 8 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0141] In one embodiment, a computer device is provided, including a memory and a processor, wherein a computer program is stored in the memory, and when the processor executes the computer program, the following steps are implemented:

[0142] Detecting voice information in the target user's environment when playing the target voice data;

[0143] If the voice information is the voice information of the target user, pausing the playing of the target voice data and recording the playing position of the target voice data at the pausing time;

[0144] Identify the semantics of the target user's voice information and perform a response operation corresponding to the semantics;

[0145] After the response operation is completed, the remaining content of the target voice data is played from the playback position at the pause time.

[0146] The above embodiment provides a computer-readable storage medium, whose implementation principle and technical effect are similar to those of the above method embodiment, and will not be repeated here.

[0147] In one embodiment, a computer readable storage medium is provided, on which a computer program is stored, and when the computer program is executed by a processor, the following steps are implemented:

[0148] Detecting voice information in the target user's environment when playing the target voice data;

[0149] If the voice information is the voice information of the target user, pausing the playing of the target voice data and recording the playing position of the target voice data at the pausing time;

[0150] Identify the semantics of the target user's voice information and perform a response operation corresponding to the semantics;

[0151] After the response operation is completed, the remaining content of the target voice data is played from the playback position at the pause time.

[0152] The above embodiment provides a computer-readable storage medium, whose implementation principle and technical effect are similar to those of the above method embodiment, and will not be repeated here.

[0153] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. As an illustration and not limitation, RAM is available in many forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).

[0154] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0155] The above-mentioned embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the invention patent. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the patent of the present application shall be subject to the attached claims.

Claims

1. A human-computer interaction method, characterized in that: The method comprises: Detecting voice information in the target user's environment when playing the target voice data; Identify whether the voice information is the voice information of the target user by using a preset voice feature detection model; the preset voice feature detection model is obtained by iteratively training the initial voice feature detection model using multiple sample voice features; If the voice information is the voice information of the target user, pausing the playing of the target voice data and recording the playing position of the target voice data at the pause time; if the voice information is the voice information of a non-target user, not responding to the voice information and continuing to play the target voice data; Recognize the semantics of the voice information of the target user, and perform a response operation corresponding to the semantics; After the response operation is completed, the remaining content of the target voice data is played from the playback position at the pause time.

2. The method according to claim 1, characterized in that The identifying the semantics of the target user's voice information and performing a response operation corresponding to the semantics includes: Determining the information type of the voice information according to the semantics of the voice information of the target user; According to the information type, a response operation corresponding to the semantics is performed.

3. The method according to claim 2, characterized in that The performing a response operation corresponding to the semantics according to the information type includes: If the information type is an inquiry type, match the reply content corresponding to the semantics of the voice information of the target user from a preset knowledge base, and play the reply content; the knowledge base includes reply content corresponding to multiple information types; or, If the information type is an echo type, the remaining content of the target voice data is played from the playback position at the pause time after pausing for a preset time.

4. The method according to claim 3, characterized in that The information type being the inquiry type indicates that the target user has questions about the played voice content.

5. The method according to claim 3, characterized in that: The information type being the echo type indicates that the voice information sent by the target user is an echo voice.

6. The method according to any one of claims 1 to 5, characterized in that: The recording of the playback position of the target voice data at the pause time includes: Determine a sentence corresponding to a pause moment in the target speech data; The position of the sentence corresponding to the pause moment in the target voice data is determined as the playback position of the target voice data at the pause moment.

7. The method according to claim 6, characterized in that The training process of the sound feature detection model includes: Acquire a plurality of sample voice features and a user identifier corresponding to each of the sample voice features; Through a deep learning algorithm, the correspondence between each of the voice features and the corresponding user identifier is learned to iteratively train the initial sound feature detection model until the change amplitude of the value of the loss function of the sound feature detection model is less than a preset threshold, thereby obtaining the sound feature detection model.

8. A human-computer interaction device, characterized in that: The device comprises: A detection module, used to detect voice information in the target user's environment when playing the target voice data; The stop module identifies whether the voice information is the voice information of the target user through a preset sound feature detection model; the preset sound feature detection model is obtained by iteratively training the initial sound feature detection model using multiple sample voice features; if the voice information is the voice information of the target user, pausing the playing of the target voice data and recording the playing position of the target voice data at the pause time; if the voice information is the voice information of a non-target user, no response is made to the voice information and the target voice data continues to be played; A response module, used for recognizing the semantics of the voice information of the target user and performing a response operation corresponding to the semantics; The playing module is used to play the remaining content of the target voice data from the playing position at the pause time after the response operation is completed.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 7 are implemented.