CN117076623A - Session processing method, session processing device, electronic equipment and computer readable storage medium - Google Patents

Abstract

This application provides a conversation processing method, device, electronic equipment and computer-readable storage medium; the method includes: obtaining the inquiry content input by the conversation object in the session; performing intent recognition processing on the inquiry content to obtain the inquiry intention; according to the inquiry intention Determine the corresponding response content, and perform response processing in the session based on the response content; control the session based on the number of response processing for the same query intention in the session. Through this application, conversation efficiency can be improved and the situation of endless loop conversations can be avoided.

Description

Session processing method, device, electronic device and computer-readable storage medium

Technical field

The present application relates to the field of computer technology, and specifically to a session processing method, device, electronic device and computer-readable storage medium.

Background technique

Artificial Intelligence (AI) is a theory, method, technology and application system that uses digital computers or machines controlled by digital computers to simulate, extend and expand human intelligence, perceive the environment, acquire knowledge and use knowledge to obtain the best results. Human-Machine Conversation is an important branch of artificial intelligence. It refers to the technology that allows machines to understand and use natural language to achieve human-machine communication. It has now been widely used in customer service, navigation, shopping guidance and other scenarios.

However, in related technologies, feedback to the user's inquiry content is usually based on a fixed process. If the user repeatedly repeats similar words, the same content will be repeatedly fed back, and the entire process will fall into an infinite loop, resulting in conversational inefficiency. Low.

Contents of the invention

Embodiments of the present application provide a conversation processing method, device, electronic device and computer-readable storage medium, which can accurately identify whether the user is repeatedly asking, thereby improving conversation efficiency.

The technical solution of the embodiment of this application is implemented as follows:

The embodiment of this application provides a session processing method, including:

Get the query content entered by the session object in the session;

Perform intention recognition processing on the inquiry content to obtain the inquiry intention;

Determine the corresponding response content according to the inquiry intention, and perform response processing in the session according to the response content;

The session is controlled based on the number of times the response processing is performed for the same inquiry intention in the session.

An embodiment of the present application provides a session processing device, including:

The acquisition module is used to obtain the query content entered by the session object in the session;

An intent recognition module, used to perform intent recognition processing on the query content to obtain the query intent;

A response module, configured to determine the corresponding response content according to the inquiry intention, and to perform response processing in the session according to the response content;

A control module, configured to control the session according to the number of times of response processing for the same inquiry intention in the session.

An embodiment of the present application provides an electronic device, including:

Memory, used to store executable instructions;

The processor is configured to implement the session processing method provided by the embodiment of the present application when executing executable instructions stored in the memory.

Embodiments of the present application provide a computer-readable storage medium that stores executable instructions for causing a processor to implement the session processing method provided by embodiments of the present application when executed.

Embodiments of the present application provide a computer program product or computer program. The computer program product or computer program includes executable instructions, and the executable instructions are stored in a computer-readable storage medium. The processor of the electronic device reads the executable instruction from the computer-readable storage medium, and the processor executes the executable instruction, so that the electronic device executes the session processing method described above in the embodiment of the present application.

The embodiments of this application have the following beneficial effects:

The embodiment of the present application is applied to the conversation process. First, the query content input by the conversation object in the conversation is obtained, the query content is subjected to intent recognition processing to obtain the query intent, the corresponding response content is determined based on the query intent, and the response is made in the conversation based on the response content. processing to provide feedback on the inquiry content. The conversation is controlled based on the number of responses to the same query intention during the conversation. By controlling the conversation instead of letting it fall into an infinite loop of dialogue, conversation efficiency can be effectively improved.

Description of the drawings

In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings needed to be used in the description of the embodiments will be briefly introduced below. Obviously, the drawings in the following description are only some embodiments of the present application. For those skilled in the art, other drawings can also be obtained based on these drawings without exerting creative efforts.

Figure 1 is a schematic structural diagram of a session processing system provided by an embodiment of the present application;

Figure 2 is a schematic structural diagram of a terminal device provided by an embodiment of the present application;

Figure 3 is a schematic flowchart of the session processing method provided by the embodiment of the present application;

Figure 4 is a schematic flow chart of the session processing method provided by the embodiment of the present application;

Figure 5 is a schematic flow chart of exception control provided by the embodiment of the present application;

Figure 6 is a schematic flow chart of exception control provided by the embodiment of the present application.

Detailed ways

In order to make the purpose, technical solutions and advantages of the present application clearer, the present application will be described in further detail below in conjunction with the accompanying drawings. The described embodiments should not be regarded as limiting the present application. Those of ordinary skill in the art will not make any All other embodiments obtained under the premise of creative work belong to the scope of protection of this application.

In the following description, references to "some embodiments" describe a subset of all possible embodiments, but it is understood that "some embodiments" may be the same subset or a different subset of all possible embodiments, and Can be combined with each other without conflict. In the following description, the term "plurality" refers to at least two.

Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the technical field to which this application belongs. The terms used herein are only for the purpose of describing the embodiments of the present application and are not intended to limit the present application.

In related technologies, feedback to the user's inquiry content is usually based on a fixed process. If the user repeats similar words repeatedly, the same content will be fed back repeatedly, and the entire process will fall into an infinite loop.

For example, in an intelligent voice training scenario, the administrator of the training system will preset the speaking process, and students need to follow the system prompts for voice training. If the student discovers that a certain sentence is said, the system will definitely react in the same way. At this time, the student may repeat the sentence repeatedly, causing the process to repeat. This wastes time and also puts the process into an infinite loop.

In summary, the solutions provided by related technologies will lead to low conversation efficiency and fail to effectively guide users.

Embodiments of the present application provide a session processing method, device, electronic device, and computer-readable storage medium, which can improve session efficiency, effectively guide session objects, and reduce resource waste. The following describes exemplary applications of the electronic devices provided by the embodiments of the present application. The electronic devices provided by the embodiments of the present application can be implemented as notebook computers, tablet computers, desktop computers, set-top boxes, mobile devices (for example, mobile phones, portable music players, Various types of terminal devices such as personal digital assistants, dedicated messaging devices, portable gaming devices) can also be implemented as servers.

Referring to Figure 1, Figure 1 is an architectural schematic diagram of the session processing system 100 provided by the embodiment of the present application. The terminal device 400 is connected to the server 200 through the network 300, and the server 200 is connected to the database 500. The network 300 can be a wide area network or a local area network, or It's a combination of the two.

In some embodiments, taking the electronic device as a terminal device as an example, the session processing method provided in the embodiments of this application can be implemented by the terminal device. Taking the conversation object as a user, for example, the user can start a human-machine conversation in the client 410 of the terminal device 400 . The terminal device 400 obtains the inquiry content entered by the user in the session; performs intent recognition processing on the inquiry content to obtain the inquiry intention; determines the corresponding response content according to the inquiry intention, and performs response processing in the session according to the response content; and responds to the same inquiry in the session. The number of times the intention is to perform response processing and control the session. This scenario can support offline human-computer dialogue and is suitable for situations where there is no network environment, poor network status, or no network required.

In some embodiments, the session processing method provided by the embodiments of this application can also be implemented by a combination of a terminal device and a server. For example, the user can start a human-computer conversation in the client 410 of the terminal device 400. The terminal device 400 obtains the query content input by the user in the session, and sends the query content to the server 200 . The server 200 performs intent recognition processing on the query content to obtain the query intent; determines the corresponding response content based on the query intent, and sends the response content to the terminal device 400, and the terminal device 400 performs response processing in the session based on the response content. The server 200 performs control processing on the session according to the number of times of response processing for the same query intention in the session. The server 200 can send the control information to the terminal device 400, and the terminal device 400 performs corresponding operations. For example, the server 200 will prompt The information is sent to the terminal device 400 and displayed by the terminal device 400 in the session. This scenario can use the computing power of the server to improve the efficiency of response and exception handling.

In some embodiments, the terminal device 400 and/or the server 200 can implement the session processing method provided by the embodiments of the present application by running a computer program. For example, the computer program can be a native program or software module in the operating system; it can be a local program. (Native) application (APP, Application), that is, a program that needs to be installed in the operating system to run, such as an application dedicated to human-machine dialogue (corresponding to the client 410 above); it can also be a small program, which only It can be a program that needs to be downloaded to the browser environment and then run; it can also be a small program or plug-in that can be embedded in any APP, where the small program or plug-in can be run or closed under the user's control. In summary, the computer program described above can be any form of application, module or plug-in.

In some embodiments, the server 200 may be an independent physical server, a server cluster or a distributed system composed of multiple physical servers, or may provide cloud services, cloud databases, cloud computing, cloud functions, cloud storage, Cloud servers for basic cloud computing services such as network services, cloud communications, middleware services, domain name services, security services, network acceleration services (Content Delivery Network, CDN), and big data and artificial intelligence platforms, among which the cloud service can be session processing Service for the terminal device 400 to call. The terminal device 400 can be a smart phone, a tablet computer, a notebook computer, a desktop computer, a smart TV, a smart watch, etc., but is not limited thereto. The terminal device and the server can be connected directly or indirectly through wired or wireless communication methods, which are not limited in the embodiments of this application.

Taking the electronic device provided by the embodiment of the present application as a terminal device as an example, it can be understood that for the case where the electronic device is a server, parts of the structure shown in Figure 2 (such as user interface, presentation module and input processing module) Can be defaulted. Referring to Figure 2, Figure 2 is a schematic structural diagram of a terminal device 400 provided by an embodiment of the present application. The terminal device 400 shown in Figure 2 includes: at least one processor 410, a memory 450, at least one network interface 420 and a user interface 430. The various components in the terminal device 400 are coupled together via a bus system 440 . It can be understood that the bus system 440 is used to implement connection communication between these components. In addition to the data bus, the bus system 440 also includes a power bus, a control bus, and a status signal bus. However, for the sake of clarity, the various buses are labeled bus system 440 in FIG. 2 .

The processor 410 may be an integrated circuit chip with signal processing capabilities, such as a general-purpose processor, a digital signal processor (DSP), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware. Components, etc., wherein the general processor can be a microprocessor or any conventional processor, etc.

User interface 430 includes one or more output devices 431 that enable the presentation of media content, including one or more speakers and/or one or more visual displays. User interface 430 also includes one or more input devices 432, including user interface components that facilitate user input, such as a keyboard, mouse, microphone, touch screen display, camera, and other input buttons and controls.

Memory 450 may be removable, non-removable, or a combination thereof. Exemplary hardware devices include solid state memory, hard disk drives, optical disk drives, etc. Memory 450 optionally includes one or more storage devices physically located remotely from processor 410 .

Memory 450 includes volatile memory or non-volatile memory, and may include both volatile and non-volatile memory. The non-volatile memory may be a read-only memory (ROM, Read Only Memory), and the volatile memory may be a random-access memory (RAM, Random Access Memory). The memory 450 described in the embodiments of this application is intended to include any suitable type of memory.

In some embodiments, the memory 450 is capable of storing data to support various operations, examples of which include programs, modules, and data structures, or subsets or supersets thereof, as exemplarily described below.

The operating system 451 includes system programs used to process various basic system services and perform hardware-related tasks, such as the framework layer, core library layer, driver layer, etc., which are used to implement various basic services and process hardware-based tasks;

Network communications module 452 for reaching other computing devices via one or more (wired or wireless) network interfaces 420, example network interfaces 420 include: Bluetooth, Wireless Compliance Certified (WiFi), and Universal Serial Bus ( USB, Universal Serial Bus), etc.;

Presentation module 453 for enabling the presentation of information (e.g., a user interface for operating peripheral devices and displaying content and information) via one or more output devices 431 (e.g., display screens, speakers, etc.) associated with user interface 430 );

An input processing module 454 for detecting one or more user inputs or interactions from one or more input devices 432 and translating the detected inputs or interactions.

In some embodiments, the session processing device provided by the embodiment of the present application can be implemented in software. Figure 2 shows the session processing device 455 stored in the memory 450, which can be software in the form of programs, plug-ins, etc., including the following Software modules: acquisition module 4551, intention recognition module 4552, response module 4553 and control module 4554. These modules are logical, so they can be combined or further split in any way according to the functions implemented. The functions of each module are explained below.

The session processing method provided by the embodiment of the present application will be described with reference to the exemplary application and implementation of the electronic device provided by the embodiment of the present application.

Referring to Figure 3, Figure 3 is a schematic flowchart of a session processing method provided by an embodiment of the present application, which will be described in conjunction with the steps shown in Figure 3.

In step 101, the query content entered by the conversation object in the conversation is obtained.

Here, a session is established with the session object, and the query content entered by the session object in the session is obtained.

The embodiment of this application does not limit the type of session object. It can be a real user or a non-user, such as a script program.

The embodiment of the present application does not limit the form of the inquiry content, which may be in a multimedia form, that is, including at least one of text, image, and voice. The same applies to the response content below. In some embodiments, the query content may be a plain text message entered by the user.

In step 102, intent recognition processing is performed on the inquiry content to obtain the inquiry intention.

After obtaining the inquiry content, perform intention recognition processing on the inquiry content to obtain the inquiry intention. The embodiment of the present application does not limit the method of intention recognition processing. For example, a keyword library corresponding to each inquiry intention can be established (for example, the keywords corresponding to the inquiry intention "say hello" can include "hello", "hello", etc.). And perform keyword recognition processing on the query content. If the identified keywords fall into a certain keyword library, the query intention corresponding to the keyword library will be used as the query intention corresponding to the query content; for another example, the intention can be pre-trained Recognition model, and perform intent recognition processing on the query content through the trained intent recognition model to obtain the query intent. The type of the intent recognition model is not limited, for example, it can be a convolutional neural network (CNN) model. wait.

In step 103, the corresponding response content is determined based on the query intention, and response processing is performed in the session based on the response content.

After determining the inquiry intention corresponding to the inquiry content, the corresponding response content is determined according to the inquiry intention, and response processing is performed in the session according to the response content, for example, the response content is displayed in the session interface to realize human-computer dialogue.

In some embodiments, the above-mentioned determination of the corresponding response content based on the query intention can be implemented in this way: performing any of the following processes: generating response content corresponding to the query intention through a response model; Query processing is performed in the query to obtain the response content; wherein the utterance library includes at least one query intention and at least one response content corresponding to each query intention.

Here, two methods are provided to determine the response content.

The first method is to use a pre-trained response model to generate response content corresponding to the query intention. This method can enhance the randomness of the response content and improve the user experience.

The second method is to establish a vocabulary library in advance. The vocabulary library includes at least one inquiry intention and at least one response content corresponding to each inquiry intention. For example, for the inquiry intention "Greetings", the corresponding response content in the vocabulary library includes "Hello, what can I do for you?" and "Hello, what do you need from me?". This method can improve the response speed and achieve rapid response.

In step 104, the session is controlled according to the number of times of response processing for the same inquiry intention in the session.

The above-mentioned steps 101 to 103 may be executed multiple times, that is, the electronic device and the conversation partner have an interactive dialogue in the conversation. Each time the conversation partner inputs query content, the electronic device performs a response process.

However, in some cases, repetitive conversations may occur within a conversation, such as conversation subjects repeatedly typing similar queries. Therefore, in the embodiment of the present application, the number of times of response processing for the same inquiry intention in the conversation is determined. Based on this number, it can be determined whether the conversation object is suspected of repeated inquiry based on the same inquiry intention, and then the conversation is processed accordingly. Control processing, which may include at least one of closing the session, prompting, and continuing to perform step 101.

It is worth noting that the embodiment of the present application does not limit the timing of counting "the number of times a response is processed for the same query intention in the session". For example, the electronic device can count every time a response is processed, or it can Statistics are performed every preset period (such as one minute).

In some embodiments, the number of times of response processing for the same inquiry intention in a session includes any of the following: the number of consecutive response processes for the same inquiry intention; the cumulative number of response processes for the same inquiry intention.

Here, the number of times of response processing for the same inquiry intention may be the number of consecutive response processes or the number of cumulative response processes, which will be described separately below.

The first is the number of consecutive response processes, which requires that response processes for the same inquiry intention must appear continuously. For example, if the conversation partner continuously inputs three pieces of inquiry content, all of which correspond to inquiry intention A, then the electronic device determines that the number of consecutive response processes for inquiry intention A is 3 after responding to the last inquiry content. For another example, after the conversation partner continuously inputs two query contents corresponding to query intention A, and then inputs one query content corresponding to query intention B, the electronic device responds to the query content for the query intention after processing the response to the last query content. The number of consecutive response processes performed by A is cleared. After that, if the conversation partner inputs another inquiry content corresponding to inquiry intention B, the electronic device determines the number of continuous response processes for inquiry intention B after performing the corresponding response processing. is 2. Continuous response processing has high requirements for continuity and is suitable for scenarios with high requirements for supervision of repeated conversations.

The second type is the cumulative number of response processes, which does not require that response processes for the same inquiry intention occur continuously, that is, they can be continuous or discontinuous. For example, after the conversation partner continuously inputs two query contents corresponding to query intention A, then inputs one query content corresponding to query intention B, and then inputs another query content corresponding to query intention A, the electronic device will process the last query content. After the inquiry content is responded to, it is determined that the cumulative number of response processes for inquiry intention A is 3 times, and the cumulative number of response processes for inquiry intention B is 1 time. For another example, after the conversation partner continuously inputs two query contents corresponding to query intention A, and then continuously inputs two query contents corresponding to query intention B, the electronic device determines the query intention after responding to the last query content. The number of cumulative response processes performed by A is 2 times, and the cumulative number of response processes performed by B for the inquiry intention is 2 times. This applies to scenarios with broader requirements for monitoring repeated conversations.

In some embodiments, when performing response processing in the session according to the response content, it also includes: counting intention tags corresponding to the inquiry intention; wherein different inquiry intentions correspond to different intention tags; The recorded value determines the number of times in the session that the query intention corresponding to the intention tag is responded to.

In order to facilitate the counting of the number of response processing, for each query intention, a corresponding intention tag can be established in advance. The intention tag is in the form of a numerical value, equivalent to a counter, and can be initially zero. Different query intentions correspond to different intention tags.

When the electronic device recognizes that the inquiry content belongs to a certain inquiry intention (taking inquiry intention A as an example) and performs response processing, it counts the intention tags corresponding to inquiry intention A, such as performing an increment operation. In this way, for inquiry intention A, the number of times of response processing for inquiry intention A in the session can be determined by the value recorded in the corresponding intention tag. Of course, in some cases, the clearing of intent tags may also be involved, such as limiting the number of consecutive response processing times. Through the above method, the convenience and real-time performance of counting the number of response processing can be improved.

In some embodiments, step 104 can be further restricted. For example, the number of response processes for the same inquiry intention in the session can be restricted to a specific time period (such as the last minute) or the last L response processes, where , L is an integer greater than 1.

Through the embodiment shown in Figure 3, counting the number of response processing for the same inquiry intention in the conversation and performing corresponding control processing can effectively avoid the situation of the conversation falling into an infinite loop, improve the conversation efficiency, and at the same time improve the efficiency of the electronic device during the conversation. The actual utilization of the computing resources consumed.

In some embodiments, see Figure 4, which is a schematic flowchart of the session processing method provided by the embodiment of the present application. Step 104 shown in Figure 3 can be implemented through step 201 and step 202.

In step 201, when the number of response processes for the same query intention in the session is greater than the number threshold, exception control is performed on the session.

Here, when the number of response processing for the same inquiry intention in the conversation is greater than the number threshold, it is determined that the conversation object is suspected of repeated inquiry, and at this time, the conversation is abnormally controlled. The exception control may include at least one of prompting (an exception prompt or a guidance prompt) and closing the session. After the prompting, the session can be closed, or the inquiry content of the session object can be continued to be obtained.

In step 202, when the number of response processing for each inquiry intention in the session is less than or equal to the number threshold, continue to perform the step of obtaining the inquiry content input by the session object in the session.

Here, when the number of response processing for each inquiry intention in the conversation is less than or equal to the number threshold, it is determined that the conversation object is not suspected of repeated inquiry, and then the normal conversation can be continued, that is, step 101 is continued. .

Through the embodiment shown in FIG. 4 , differentiated processing is performed based on the relationship between the number of response processing times and the number threshold, which can improve the pertinence of session processing.

In some embodiments, see FIG. 5 , which is a schematic flow chart of exception control provided by an embodiment of the present application. Step 201 shown in FIG. 4 can be implemented through steps 301 to 303 .

In step 301, when the number of response processing for the same inquiry intention in the session is greater than the number threshold, the inquiry similarity between multiple target inquiry contents is determined; wherein the target inquiry content represents the response to the same inquiry. Intention to respond to the query content answered.

When the number of response processing for the same inquiry intention in a conversation is greater than the number threshold, it proves that the conversation object may be suspected of repeated inquiries. In the embodiment of the present application, further confirmation can be made from the perspective of the conversation object. First, the query similarity between multiple target query contents is calculated. The query similarity represents the degree of similarity between the multiple target query contents, wherein the target query content represents the response to the same query intention. content of the inquiry.

In step 302, when the query similarity is greater than the first similarity threshold, an exception prompt is provided to the conversation object to prompt to stop inputting query content corresponding to the same query intention.

Here, when the query similarity is greater than the first similarity threshold, it is confirmed that a repeated query is constituted. At this time, an exception prompt is given to the conversation partner to prompt to stop inputting the query content corresponding to the above-mentioned same query intention.

It is worth noting that the embodiment of the present application does not limit the method of prompting. For example, relevant prompts can be displayed in the conversation interface, and conversation objects can be prompted through other methods (such as emails, etc.).

In step 303, when the query similarity is less than or equal to the second similarity threshold, a guidance prompt is provided to the conversation object to prompt the input of query content corresponding to other query intentions; wherein the other query intentions are different from the same query intention; first The similarity threshold is greater than or equal to the second similarity threshold.

When the query similarity is less than or equal to the second similarity threshold, it is confirmed that it does not constitute a repeated query. It may be that the response processing times are too many due to an error in intent recognition or other reasons. At this time, a guidance prompt is provided to the conversation object to prompt Enter the query content corresponding to other query intentions; where the other query intentions are different from the same query intention mentioned above.

The first similarity threshold is greater than or equal to the second similarity threshold, and can be set according to the actual situation. For example, the first similarity threshold can be set to be equal to the second similarity threshold.

In some embodiments, the above-mentioned determination of query similarity between multiple target query contents can be implemented in this way: performing any one of the following processes: determining the similarity between adjacent target query contents as Query similarity; determine the similarity between adjacent target query contents, and fuse multiple similarities to obtain query similarity.

Here, when there are only two target query contents, the similarity between the two target query contents can be directly used as the query similarity.

When the number of target query contents is more than three, you can first determine the similarity between two adjacent target query contents. In this way, multiple similarities will be obtained, and then each similarity will be regarded as the query similarity. Since the number of query similarities includes multiple, "the query similarity is greater than the first similarity threshold" in step 302 may mean that all query similarities are greater than the first similarity threshold, or it may refer to a preset number/ The query similarity of the preset ratio is greater than the first similarity threshold, and the preset number/preset ratio can be set according to the actual application scenario. The same applies to step 303 and step 304 in FIG. 6 .

When the number of target query contents is more than three, you can also first determine the similarity between two adjacent target query contents. In this way, multiple similarities will be obtained, and then the multiple similarities will be fused to obtain The result is used as query similarity. The fusion process may be an arithmetic average or a weighted average, etc., and is not limited to this. In this way, it is only necessary to compare the obtained query similarity with the first similarity threshold and the second similarity threshold.

Through the above method, the flexibility of threshold comparison is improved, and the calculation method of query similarity can be selected according to the needs in actual scenarios.

Through the embodiment shown in Figure 5, it is possible to further confirm whether a repeated inquiry is constituted, identify the root cause of repeated conversations, and then carry out targeted processing to improve the accuracy and flexibility of processing.

In some embodiments, see Figure 6, which is a schematic flow chart of exception control provided by an embodiment of the present application. After step 301 shown in Figure 5, in step 304, when the query similarity is less than or When equal to the first similarity threshold and greater than the second similarity threshold, continue to perform the step of obtaining the query content input by the conversation object in the conversation.

Here, when the first similarity threshold is greater than the second similarity threshold, a third situation can also be extended, that is, the query similarity is less than or equal to the first similarity threshold and greater than the second similarity threshold. In this case , determining that the conversation object may need to be confirmed through multiple inquiries due to various reasons (such as being unable to hear or understand the response content), therefore, the conversation object will not be disturbed (no prompts will be given), but the normal conversation will continue, that is, continue Get the query content newly entered by the session object in the session.

Through the embodiment shown in Figure 6, new situations of dialogue are further expanded and the flexibility of conversation processing is improved.

Below, an exemplary application of the embodiments of the present application in a practical application scenario will be described, that is, the electronic device implements human-computer interaction through a human-computer dialogue system, thereby answering user questions and guiding the user. For ease of understanding, instructions will be given in step-by-step format.

1) Unified mapping of corresponding speech skills of the system.

During the conversation, the utterance content of the human-computer dialogue system's response (corresponding to the response content above) corresponds to the user's intention (inquiry intention). Although the response utterance is not necessarily exactly the same, it must be intelligent based on the user's intention. Obtained by generating or calling the vocabulary library.

In the embodiment of the present application, no matter which way the human-machine dialogue system obtains the response utterances, it can mark the utterances returned with the same intention. For example: when the user says "Hello" in the conversation (corresponding to the query content above), the human-computer dialogue system will reply "Hello, what can I do for you?" or "Hello, what do you need from me?" ", the actual meaning of these two replies is the same, they are both responses to the intention of greeting "Hello". Therefore, replies with the same intention can be labeled with a common label (corresponding to the intention label above), and this common label is the discourse mapping. In the above example, the user inputs "Hello" with the intention of saying hello, then the system's responses to this intention will all have a common label of "hello".

2) System reply content sequence and frequency detection.

During the human-machine dialogue, each response of the system can be recorded in sequence. Counting starts from the second system reply. If the label of the current system reply is consistent with the label of the previous reply, the number of occurrences of the label will be increased by one; if the label of the current system reply is inconsistent with the label of the previous reply, the label will be cleared. number of occurrences.

Whenever the number of occurrences of a tag is increased by one, the human-machine dialogue system determines whether the number of occurrences of the tag (corresponding to the number of consecutive response processing for the same query intention above) reaches the threshold N (corresponding to the number of times threshold above). If it reaches the threshold N, it is considered that the user is suspected of repeated inquiries, and the abnormal process is triggered (corresponding to the exception control above); if the threshold N is not reached, the abnormal process is not triggered. Among them, N is an integer greater than 1.

3) Abnormal process.

Since the user seems to have made repeated inquiries, further confirmation is made from the perspective of user input in this step. The human-computer dialogue system performs similarity detection on the N pieces of content recently input by the user. For example, it calculates the similarity between two adjacent pieces of content among the N pieces of content recently input by the user. If all similarities are greater than the threshold K (corresponding to The first similarity threshold above (in order to prevent the user from repeating almost the same words, so the K value here needs to be set higher), then it is deemed that the user has repeated it intentionally. At this time, the human-computer dialogue system will give an abnormal prompt to the user. , such as "Do not repeat similar words continuously"; if all similarities are less than or equal to the threshold M (corresponding to the second similarity threshold above, in order to prevent the human-machine dialogue system from identifying the wrong user intention, causing the human-machine dialogue system to Continuously returning the same content, the M value needs to be set lower), it is determined that the user's intention has been recognized incorrectly. At this time, the human-computer dialogue system will provide guidance prompts to the user to prompt the user to enter the content of other intentions, that is, to open a new Process dialogue, such as "Oh, the system is tired of replying, let's talk about something else"; if it is other situations, consider that in real scenarios, users may ask repeatedly due to unclear hearing, incomprehension, etc. , at this time, the human-machine dialogue system will not prompt.

Through the above examples, the probability of possible repetition of the dialogue process is greatly reduced, the user experience of human-computer dialogue is improved, and the error process is effectively guided.

The following continues to describe an exemplary structure in which the session processing device 455 provided by the embodiment of the present application is implemented as a software module. In some embodiments, as shown in Figure 2, the software module stored in the session processing device 455 of the memory 450 may include: The acquisition module 4551 is used to obtain the inquiry content input by the conversation object in the session; the intention identification module 4552 is used to perform intention identification processing on the inquiry content and obtain the inquiry intention; the response module 4553 is used to determine the corresponding response content according to the inquiry intention. , and perform response processing in the session according to the response content; the control module 4554 is used to perform control processing on the session according to the number of response processing for the same inquiry intention in the session.

In some embodiments, the control module 4554 is also configured to: perform abnormal control on the session when the number of response processing for the same query intention in the session is greater than the number threshold; when the number of response processing for each query in the session When the number of times that the response processing is intended is less than or equal to the number of times threshold, continue to perform the step of obtaining the query content input by the conversation object in the conversation.

In some embodiments, the control module 4554 is also used to: determine the query similarity between multiple target query contents; wherein the target query content represents the query content responded to by responding to the same query intention; When the query similarity is greater than the first similarity threshold, an exception prompt is given to the conversation object to prompt to stop inputting query content corresponding to the same query intention; when the query similarity is less than or equal to the second similarity When the threshold is reached, a guidance prompt is provided to the conversation object to prompt the input of inquiry content corresponding to other inquiry intentions; wherein the other inquiry intentions are different from the same inquiry intention; wherein the first similarity threshold is greater than or equal to The second similarity threshold.

In some embodiments, the first similarity threshold is greater than the second similarity threshold; the control module 4554 is also configured to: when the query similarity is less than or equal to the first similarity threshold and greater than the third similarity threshold, When the similarity threshold is 2, continue to perform the step of obtaining the query content input by the conversation object in the conversation.

In some embodiments, the control module 4554 is also configured to: perform any of the following processes: determine the similarity between adjacent target query contents as the query similarity; determine the similarity between adjacent target query contents. similarity, and perform fusion processing on multiple similarities to obtain the query similarity.

In some embodiments, the session processing device 455 also includes a tag counting module, configured to: when performing response processing in the session according to the response content, count the intention tags corresponding to the query intention; wherein, Different inquiry intentions correspond to different intention tags; according to the value recorded in the intention tag, the number of times of response processing for the inquiry intention corresponding to the intention tag is determined in the session.

In some embodiments, the number of times of response processing for the same inquiry intention in a session includes any of the following: the number of consecutive response processes for the same inquiry intention; the cumulative number of response processes for the same inquiry intention.

Embodiments of the present application provide a computer program product or computer program. The computer program product or computer program includes executable instructions, and the executable instructions are stored in a computer-readable storage medium. The processor of the electronic device reads the executable instruction from the computer-readable storage medium, and the processor executes the executable instruction, so that the electronic device executes the session processing method described above in the embodiment of the present application.

Embodiments of the present application provide a computer-readable storage medium storing executable instructions. The executable instructions are stored therein. When the executable instructions are executed by a processor, they will cause the processor to execute the method provided by the embodiments of the present application, such as , the session processing method shown in Figure 3.

In some embodiments, the computer-readable storage medium may be a memory such as FRAM, ROM, PROM, EPROM, EEPROM, flash memory, magnetic surface memory, optical disk, or CD-ROM; it may also include one or any combination of the above memories. Various equipment.

In some embodiments, executable instructions may take the form of a program, software, software module, script, or code, written in any form of programming language, including compiled or interpreted languages, or declarative or procedural languages, and their May be deployed in any form, including deployed as a stand-alone program or deployed as a module, component, subroutine, or other unit suitable for use in a computing environment.

As an example, executable instructions may, but do not necessarily correspond to, files in a file system and may be stored as part of a file holding other programs or data, for example, in a Hyper Text Markup Language (HTML) document. in one or more scripts, stored in a single file specific to the program in question, or in multiple collaborative files (e.g., files storing one or more modules, subroutines, or portions of code).

As examples, executable instructions may be deployed to execute on one computing device, or on multiple computing devices located at one location, or alternatively, on multiple computing devices distributed across multiple locations and interconnected by a communications network execute on.

The above are only examples of the present application and are not used to limit the protection scope of the present application. Any modifications, equivalent substitutions and improvements made within the spirit and scope of this application are included in the protection scope of this application.

Claims (10)

1. A session processing method, comprising:

acquiring inquiry content input by a session object in a session;

performing intention recognition processing on the query content to obtain a query intention;

determining corresponding response content according to the query intention, and performing response processing in the session according to the response content;

and according to the times of response processing aiming at the same query intention in the session, controlling the session.

2. The method according to claim 1, wherein the controlling the session according to the number of times of response processing for the same query intention in the session includes:

When the number of times of response processing aiming at the same query intention in the session is larger than a threshold value of times, abnormal control is carried out on the session;

and when the number of times of response processing for each query intention in the session is smaller than or equal to the number threshold, continuing to execute the step of acquiring the query content input by the session object in the session.

3. The method of claim 2, wherein the performing anomaly control on the session comprises:

determining query similarity between a plurality of target query contents; wherein the target query content characterizes the query content responded by the response processing for the same query intention;

when the query similarity is greater than a first similarity threshold, carrying out abnormal prompt on the session object to prompt to stop inputting query content corresponding to the same query intention;

when the query similarity is smaller than or equal to a second similarity threshold, guiding prompt is carried out on the session object so as to prompt the input of query contents corresponding to other query intentions; wherein the other query intent is distinct from the same query intent;

wherein the first similarity threshold is greater than or equal to the second similarity threshold.

4. A method according to claim 3, wherein the first similarity threshold is greater than the second similarity threshold; after the query similarity between the plurality of target query contents is determined, the method further comprises:

and when the query similarity is smaller than or equal to the first similarity threshold and larger than the second similarity threshold, continuing to execute the step of acquiring the query content input by the session object in the session.

5. The method of claim 3, wherein determining query similarity between a plurality of target query contents comprises:

any one of the following processes is performed:

determining the similarity between adjacent target query contents to be used as query similarity;

and determining the similarity between adjacent target query contents, and carrying out fusion processing on a plurality of similarities to obtain the query similarity.

6. The method according to any one of claims 1 to 5, characterized in that when a response process is performed in the session according to the response content, the method further comprises:

counting the intention labels corresponding to the query intention; wherein different query intents correspond to different intent tags;

And determining the number of times of response processing for the query intention corresponding to the intention label in the session according to the numerical value recorded by the intention label.

7. The method according to any one of claims 1 to 5, wherein the number of times of reply processing for the same query intention in the session includes any one of:

the number of consecutive reply processes for the same interrogation intention;

the number of times of response processing is accumulated for the same query intention.

8. A session processing apparatus, comprising:

the acquisition module is used for acquiring inquiry content input by the session object in the session;

the intention recognition module is used for carrying out intention recognition processing on the inquiry content to obtain inquiry intention;

the response module is used for determining corresponding response content according to the query intention and carrying out response processing in the session according to the response content;

and the control module is used for controlling the session according to the times of response processing aiming at the same query intention in the session.

9. An electronic device, comprising:

a memory for storing executable instructions;

A processor configured to implement the session processing method according to any one of claims 1 to 7 when executing the executable instructions stored in the memory.

10. A computer readable storage medium storing executable instructions for implementing the session handling method according to any one of claims 1 to 7 when executed by a processor.