WO2020258082A1

WO2020258082A1 - Information recommendation method and apparatus, electronic device and storage medium

Info

Publication number: WO2020258082A1
Application number: PCT/CN2019/093050
Authority: WO
Inventors: 郭子亮
Original assignee: 深圳市欢太科技有限公司; Oppo广东移动通信有限公司
Priority date: 2019-06-26
Filing date: 2019-06-26
Publication date: 2020-12-30
Also published as: CN113366467A

Abstract

The embodiments of the present application disclose an information recommendation method and apparatus, an electronic device and a storage medium. Said method comprises: a voice assistant receiving voice information upon being started; identifying the voice information on the basis of a plurality of slots corresponding to a target search requirement; if a target parameter corresponding to a target slot in the plurality of slots is successfully identified from the voice information, generating a search instruction carrying target parameters respectively corresponding to the plurality of slots, the target parameter corresponding to the target slot being the parameter related to a restaurant; sending the search instruction to a target restaurant information providing platform, and receiving restaurant information returned by the target restaurant information providing platform in response to the search instruction; and displaying a card and displaying restaurant information on the card. Said method allows setting a required slot by configuring a search requirement, and then generating a subsequent search instruction on the basis of the set slot, thereby improving the accuracy of feedback for restaurant searching.

Description

Information recommendation method, device, electronic equipment and storage medium

Technical field

This application relates to the field of computer technology, and more specifically, to an information recommendation method, device, electronic equipment, and storage medium.

Background technique

With the development of voice recognition technology, more electronic devices are equipped with voice assistants to interact with users of electronic devices. In the process of interacting with voice assistants, users can use voice assistants to complete some operations. However, related In the process of voice assistant responding to users' needs, the user experience needs to be improved.

Summary of the invention

In view of the above problems, this application proposes an information recommendation method, device, electronic equipment, and storage medium to improve the above problems.

In the first aspect, this application provides an information recommendation method applied to a voice assistant. The method includes: acquiring voice information; recognizing the voice information based on multiple slots corresponding to target search requirements; When the target parameter corresponding to the target slot of the multiple slots is successfully identified in the voice information, a search instruction carrying the target parameter corresponding to each of the multiple slots is generated, and the target corresponding to the target slot is generated. The parameter is the parameter about the restaurant; the search instruction is sent to the target restaurant information providing platform, and the restaurant information returned by the target restaurant information providing platform in response to the search instruction is received; the card is displayed, and the card is displayed Restaurant information.

In a second aspect, the present application provides an information recommendation device. The device includes: a voice information receiving unit for acquiring voice information; and a voice recognition unit for responding to the voice information based on multiple slots corresponding to target search requirements. Information; the instruction generation unit is used to generate the corresponding target parameter of the multiple slots if the target parameter corresponding to the target slot of the multiple slots is successfully recognized from the voice information The target parameter search instruction, the target parameter corresponding to the target slot is the parameter about the restaurant; the restaurant information acquisition unit is used to send the search instruction to the target restaurant information providing platform, and receive the target restaurant information providing platform response The restaurant information returned by the search instruction; an information display unit for displaying a card, and displaying the restaurant information in the card.

In a third aspect, this application provides an electronic device including one or more processors and a memory; one or more programs, wherein the one or more programs are stored in the memory and configured to be The one or more processors execute the above-mentioned methods.

In a fourth aspect, the present application provides a computer-readable storage medium with program code stored in the computer-readable storage medium, wherein the above method is executed when the program code is running, and control confusion can also be avoided.

The present application provides an information recommendation method, device, electronic equipment, and storage medium that recognize the voice information in multiple slots corresponding to target search requirements; if the voice information is successfully recognized from the voice information When the target parameter corresponding to the target slot of the multiple slots is generated, a search instruction carrying the target parameter corresponding to each of the multiple slots is generated; the search instruction is sent to the target restaurant information providing platform, and the search instruction is received. The target restaurant information providing platform returns restaurant information in response to the contraction instruction; displaying a card, and displaying the restaurant information in the card. In this way, the slot required by configuring the search requirement is realized, and then the subsequent search instruction is generated based on the set slot, thereby improving the accuracy of the feedback for the restaurant search.

Description of the drawings

In order to more clearly describe the technical solutions in the embodiments of the present application, the following will briefly introduce the drawings needed in the description of the embodiments. Obviously, the drawings in the following description are only some embodiments of the present application. For those skilled in the art, other drawings can be obtained based on these drawings without creative work.

Fig. 1 shows a flowchart of an information recommendation method proposed by an embodiment of the present application;

Fig. 2 shows a schematic diagram of a voice information collection interface proposed by an embodiment of the present application;

FIG. 3 shows a schematic diagram of a parameter setting interface proposed by an embodiment of the present application;

FIG. 4 shows a schematic diagram of a result of adding a parameter voice setting interface according to an embodiment of the present application;

Fig. 5 shows a schematic diagram of a card proposed in an embodiment of the present application;

Fig. 6 shows a flowchart of an information recommendation method proposed by another embodiment of the present application;

FIG. 7 shows a schematic diagram of a voice assistant communicating with a server of a client for providing restaurant information according to an embodiment of the present application;

FIG. 8 shows a flowchart of an information recommendation method proposed by still another embodiment of the present application;

FIG. 9 shows a flowchart of an information recommendation method proposed by another embodiment of the present application;

FIG. 10 shows a structural block diagram of a voice information processing device proposed in an embodiment of the present application;

FIG. 11 shows a structural block diagram of a voice information processing device proposed by another embodiment of the present application;

FIG. 12 shows a structural block diagram of a voice information processing device proposed by still another embodiment of the present application;

FIG. 13 shows a structural block diagram of an electronic device of the present application for executing the information recommendation method according to an embodiment of the present application;

Fig. 14 is a storage unit for storing or carrying program code for implementing the information recommendation method according to the embodiment of the present application.

Detailed ways

The technical solutions in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all of the embodiments. Based on the embodiments in this application, all other embodiments obtained by those of ordinary skill in the art without creative work fall within the protection scope of this application.

With the development of voice recognition technology, electronic devices such as smart phones, tablets, and computers are equipped with intelligent voice assistants. For example, Apple's Siri, Samsung's Bixby, Google Assistant, Amazon Alex, etc. The voice assistant can be regarded as an intelligent application. The user can help the user solve some practical problems or replace the user in operating the electronic device through the intelligent interaction of the intelligent dialogue with the voice assistant and the instant question and answer. For example, in a related scenario, when the voice assistant of an electronic device detects that the user input "Help me open Baidu map", the electronic device can recognize that the user's intention is to use the Baidu map application, so that it can start Baidu The map starts. In another related scenario, when the voice assistant of the electronic device detects that the user input "Where is there a parking lot nearby", the electronic device can recognize that the user is looking for a parking lot within a certain range, then the electronic device can be based on the location Search for parking lots and display the search results.

However, the inventor found in his research on voice assistants that the related voice assistants still cannot well recognize the semantics of the restaurant-related parameters entered by the user, or even if they can respond, they cannot bring users better users. Experience. For example, in one way, after the relevant voice assistant recognizes the user input I want to eat hot pot, it either uses "hot pot" or "eat hot pot" as keywords to search for resources in a similar way to search engines, and then displays the searched , Or it may be directly searched for nearby hot pot restaurants for display. However, the inventor found that the relevant voice assistant's response to the restaurant-related parameters is not the user's real intention, resulting in an unprepared search for the user's intention, thereby failing to improve the user's experience or satisfy the user's actual needs. Furthermore, the inventor also found that in the process of directly using related shopping software for restaurant search, users need to repeatedly slide the page to find the restaurant they need, which will also cause a poor user experience.

Therefore, the inventor proposes an information recommendation method, device, electronic device, and storage medium that can improve the above-mentioned problems in this application, thereby realizing the slot required by configuring the search requirement in this way, and then based on the set slot Generate subsequent search instructions, thereby improving the accuracy of feedback for restaurant search.

The embodiments of the present application will be described in detail below with reference to the accompanying drawings.

Please refer to FIG. 1, an information recommendation method provided by an embodiment of the present application is applied to a voice assistant, and the method includes:

Step S110: Acquire voice information.

Among them, it is understandable that the voice assistant may be an application program that runs independently in an electronic device. It can also be a component configured in a certain application. During the user's use, the user can trigger the activation of the voice assistant by touching the physical buttons of the electronic device, or trigger the activation of the voice assistant by touching the virtual buttons displayed on the electronic device.

For example, as a way, when the electronic device is equipped with the HOME button, the electronic device can be configured in advance to trigger the associated target application to start by long pressing the HOME button or double-clicking the HOME button, or long pressing the HOME button or double-clicking the HOME button In the case of triggering a component in the associated target application to start, configure the associated target application or a component in the target application as a voice assistant, so that the voice can be triggered by long pressing the HOME button or double-clicking the HOME button The assistant starts. As another way, if the electronic device is equipped with a touch screen, the electronic device can be configured with an entrance that triggers the activation of the voice assistant on the desktop of the system or in an application. In the case where the voice assistant portal is configured on the desktop, the portal can be a desktop application icon named as the voice assistant, and when an portal that triggers the voice assistant to start is configured in a certain application, the portal can be a certain A control named Voice Assistant, for example, a text control or a button control.

When the voice assistant is started, the electronic device can display the interface 99 as shown in Figure 2. After the interface 99 is displayed, the electronic device will trigger the configured microphone or other physical voice collection components to start sound collection, so as to start Your voice assistant can get the collected voice information.

Step S120: Recognizing the voice information based on multiple slots corresponding to the target search requirement.

It is understandable that multiple search requirements can be configured in advance, and a corresponding slot can be configured for each search requirement in advance. For example, search requirements for restaurants may be pre-configured, and the slots corresponding to the search requirements for restaurants may include restaurant location, restaurant type, and restaurant evaluation. For another example, search requirements for travel may be pre-configured, and the slots corresponding to the search requirements for travel may include travel departure point, travel time, and travel destination. Then, when the voice assistant recognizes that the current scene is based on restaurant search, it can use the restaurant search requirement as the target search requirement, and further perform voice recognition based on the slot corresponding to the restaurant search requirement.

Among them, it is understandable that the voice information received by the voice assistant is still a voice signal, and the voice assistant also needs to convert the voice information in the form of the voice signal into the voice information in the text form, and this conversion process is to respond to the received voice information. The process of identification.

In the embodiment of the present application, the voice assistant has multiple ways to realize the recognition of the received voice information. As a way, the API (Application Programming Interface) of the third-party speech recognition system can be pre-configured. In this way, the voice assistant can wait for the third-party speech recognition system based on the pre-configured API of the third-party speech recognition system. The recognized voice signal is transmitted to a third party for recognition, and then the voice information in text form returned by the third party is received. For example, the API of a third-party speech recognition system provided by Microsoft or Google can be used.

As another way, the neural network model can be trained in advance to obtain a model capable of converting voice information in the form of sound signals into voice information in text form. Then the trained model is deployed in a designated server or service cluster, and the voice assistant can transmit the received voice information in the form of voice signals to the process of recognizing the received voice information. The server or service cluster performs recognition, and then receives the recognized text voice information returned by the server or service cluster. In another way, the voice assistant can call the API of a third-party speech recognition system for recognition, and transmit it to a designated server or service cluster for recognition and real-time selection, so as to enhance the flexibility of recognition.

In this way, optionally, the voice assistant can determine which form of recognition to perform according to the current network status. It is understandable that the communication process will be more stable when directly connected to the network through the WIFI hotspot than through the mobile communication base station. Then the voice assistant can recognize that the current access to the network is through WIFI, Call the API of the third-party speech recognition system for recognition, and when it is recognized that the mobile communication base station is currently connected to the network, it is transmitted to the designated server or service cluster for recognition. Furthermore, the voice assistant can also switch the recognition mode based on the recognition success rate. It is understandable that the pronunciation habits or speaking styles of different users will be different, so even for the same recognition method, there will be different recognition results due to the user's own pronunciation habits or speaking styles. Then in this case, the voice assistant can switch the recognition mode according to the user's own habits and combined with the recognition success rate. Among them, the voice assistant can determine that the recognition fails when it detects that the user has repeatedly recognized similar sound signals. Then the voice assistant can switch to another recognition method after recognizing a specified number of failures using a certain recognition method. For example, if the voice assistant starts to use the API of a third-party voice recognition system for recognition, and detects that it has recognized three similar sound signals in a row (the specified number of times is 3), then it is determined to switch to transmission to the server or service Identify in the cluster. It should be noted that similar sound signals can be judged by judging the error energy of the two sound signals.

It is understandable that in order to be able to recommend the required information for the user more accurately, more parameters need to be obtained as search conditions. For example, if the user wants to eat hot pot, if only the hot pot is used as a parameter to search, then only some hot pot restaurants are found in the search, and it may be any location. However, if you not only use hot pot as a parameter in the search, but also add a location (for example, the user's current location), then it can be more in line with the user's needs.

In order to facilitate more accurate recommendation, you can pre-configure the search requirements for restaurant search, and define the required slots in the search requirements. Among them, the content corresponding to each slot is the search parameter in the previous example. For example, the slots required for restaurant search may include restaurant type, restaurant rating, and restaurant location.

The voice assistant can obtain the target search requirements in a variety of ways. As a way, the target search requirement can be pre-stored locally (for example, pre-stored in the installation file of the voice assistant), then when the voice assistant recognizes that the user needs to obtain a designated restaurant before performing step S120, The target search requirement is acquired, so that when step S120 is performed, the voice information can be recognized based on the multiple slots corresponding to the acquired target search requirement.

Furthermore, as user needs change, the user's search feedback content will also change, so in order to better dynamically update the slots in the target search requirements, as a way, the voice assistant can also get updates from the network Then use the obtained updated target search requirements to replace the local target search requirements, so that the voice assistant can dynamically perform voice recognition based on the latest slots to improve the accuracy of user search feedback.

For example, the slots before the update include restaurant type, restaurant rating, and restaurant location. Then when the user says "I'm going to Chongqing to eat a hot pot with high ratings", the voice assistant can recognize that the target parameter corresponding to the slot of restaurant type is "hot pot" after voice recognition through the slot, and the slot of restaurant rating is this slot. The target parameter corresponding to the position is "high evaluation", and the target parameter corresponding to the slot of restaurant location is "Chongqing". However, as the types of hot pots are subdivided, the corresponding hot pot restaurants may also have more subdivisions. For example, hot pots will distinguish oil hot pots or butter hot pots, and the corresponding hot pot restaurants will also be classified as butter hot pots. Shop or Qingyou hot pot restaurant. So in this case, if you search for hot pot directly, users with more segmented needs may need to manually filter to search for the hot pot restaurant they need.

Then in this case, the voice assistant can detect that the user has a more segmented search request, and send the information to the information providing platform. After receiving the segmented search request, the information providing platform can automatically search The slot corresponding to the demand is updated so that the user's demand can be searched more accurately. For example, you can generate a more subdivided slot for the restaurant type slot. In this way, the information providing platform can directly transmit the updated target search requirements to the voice assistant for storage.

It should be noted that while the information providing platform is updating the slots, it can also adjust the structure of the data table storing the information, so as to better match the search requirements. For example, as shown in the following table:

IDID	餐厅类型Restaurant type	餐厅评价度Restaurant rating	餐厅位置Restaurant location	餐厅名称Restaurant name
11	火锅Hot Pot	SS	重庆Chongqing	老火锅Old hot pot

22	西餐Western food	SS	成都Chengdu	真西餐Real Western Food
33	面食pasta	SS	西安Xi'an	手工面Handmade noodles

The fields in the data table shown in the above table include primary key (ID), restaurant name, restaurant type, restaurant rating, and restaurant location. Then in this case, each slot will have a corresponding field value in the data table, so that the data can be searched accurately. Then, after the slot is updated based on the aforementioned content, the information providing platform can automatically modify the above-mentioned data table to match the new slot. As shown in the following table:

IDID	餐厅类型Restaurant type	细分类型Subdivision type	餐厅评价度Restaurant rating	餐厅位置Restaurant location	餐厅名称Restaurant name
11	火锅Hot Pot	牛油butter	SS	重庆Chongqing	老火锅Old hot pot
22	西餐Western food	法国西餐French Western Food	SS	成都Chengdu	真西餐Real Western Food
33	面食pasta	西北面食Northwest Pasta	SS	西安Xi'an	手工面Handmade noodles

As shown in the above table, in addition to the previous restaurant types, the information providing platform has added the subdivision type field to match the increased slots. Of course, in addition to subdividing the restaurant type to increase slots, more subdivided slots can also be added for other slots. For example, for the restaurant location, only the city name was originally marked. Then after the voice assistant recognizes that the user will often search for restaurants in specific areas or roads, it can add a more detailed field for the location of the restaurant, so that the location of the restaurant corresponds to the slot of the restaurant. match.

Step S130: If the target parameter corresponding to the target slot of the multiple slots is not successfully identified from the voice information, the process ends.

Step S131: If the target parameter corresponding to the target slot of the multiple slots is successfully identified from the voice information, a search instruction carrying the target parameter corresponding to each of the multiple slots is generated, so The target parameter corresponding to the target slot is a parameter about the restaurant.

As a way, the voice assistant can use multiple slots as the target slot, or only some slots or a certain slot as the target slot.

Among them, there are many ways to configure the parameters of the restaurant. As a way, several common restaurant-related parameters obtained after statistics can be configured by default when the voice assistant is installed. For example, the aforementioned "hot pot", "steak" or "western food" and so on. Furthermore, as another way, it can also be configured by users according to their own needs. As user needs change, more users expect to adopt a more personalized way to control electronic devices. For example, some users like to say "hot pot" or "noodles", etc. Then, in order to make the voice assistant adapt to users with different language expression habits, the user can be configured to configure the parameters of the restaurant.

For example, as shown in FIG. 3, the parameter setting interface 98 shown in FIG. 3 includes a parameter addition control 97 related to the restaurant and a display control 96 to which parameters related to the restaurant have been added. Furthermore, after the user clicks on the parameter adding control 97 about the restaurant, a new parameter about the restaurant can be added. In addition to adding text in the form of adding, it can also be added in the form of sound signals at the same time. For example, when the user clicks on the parameter addition control 97 about the restaurant, the user will be triggered to enter the parameter about the restaurant in the form of voice. For example, after the user input is "hot pot", in addition to the display control for the parameter about the restaurant shown in Figure 4 After 96 displays "hot pot pot" in text form, the sound signal corresponding to "hot pot pot" will be stored at the same time. Then in this way, after the voice assistant receives the voice information, in order to improve the final feedback effect, it can directly compare the received voice signal with the pre-stored voice signal. It is understandable that the voice In the signal process, the aforementioned method of comparing the signal similarity can be used to determine whether the received sound signal is similar to the pre-stored sound signal. If so, the target product information can be directly obtained without further follow-up Language to text conversion recognition, thereby improving the efficiency of information feedback.

Of course, for the restaurant-related parameters added by the user, in order to facilitate the improvement of recognition accuracy, a corresponding common name can be configured locally. For example, “hot pot” can be configured to correspond to hot pot.

Step S140: Send the search instruction to the target restaurant information providing platform, and receive restaurant information returned by the target restaurant information providing platform in response to the search instruction.

Step S150: Display a card, and display the restaurant information in the card.

As shown in FIG. 5, when the restaurant information is searched, the restaurant information can be displayed in the card 92 in the interface 99. The restaurant information may include restaurant name, restaurant icon, restaurant rating, restaurant location, etc.

Optionally, after the voice assistant displays the card and displays restaurant information, it can display more information in response to the user's touch operation, for example, display the detailed introduction of the target restaurant selected from the restaurant information, and can also display information according to the user’s Operate to make restaurant reservations.

In the information recommendation method provided by the present application, if the target parameter corresponding to the target slot of the plurality of slots is successfully recognized from the voice information, it is generated carrying the corresponding corresponding to each of the plurality of slots. The search instruction for the target parameter of the target restaurant, the search instruction is sent to the target restaurant information providing platform, and the restaurant information returned by the target restaurant information providing platform in response to the contraction instruction is received; a card is displayed, and the card is displayed Item restaurant information. In this way, the slot required by configuring the search requirement is realized, and then the subsequent search instruction is generated based on the set slot, thereby improving the accuracy of the feedback for the restaurant search.

Please refer to FIG. 6, an information recommendation method provided by an embodiment of the present application is applied to a voice assistant, and the method includes:

Step S210: Acquire voice information.

Step S220: Based on the received voice, it is determined whether it is in a closed-domain voice dialogue scene about the restaurant.

It should be noted that an open domain voice dialogue scenario and a closed domain voice dialogue scenario may be pre-configured.

Wherein, when in an open domain dialogue scenario, the voice assistant randomly selects one feedback content for feedback when recognizing that there are multiple feedback content available for responding to the received voice information. It is understandable that when in an open-domain dialogue scenario, the voice assistant has not yet determined what kind of information the user is searching for, so random information can be used for feedback. For example, when the voice assistant recognizes that the user has said "Today's weather is really good", the voice assistant can feed back information such as "Yes", "Why don't you go out for a walk", and even feedback the current weather.

When in a closed-domain voice dialogue scene about a restaurant, after completing the recognition of the voice information, if it is recognized that the slot lacks target parameters, the voice assistant will issue a prompt message to prompt the user to enter the lack of target parameters The target parameter required by the slot. It is understandable that when in a closed-domain voice dialogue scene related to a certain direction, the voice assistant has determined the information actually needed by the user, and will obtain as many parameters as possible so that the search can be performed more accurately. Among them, the voice assistant will actively send voice or text prompt information to prompt the user to input more parameters.

Among them, as a way, the voice assistant can determine whether to enter which closed domain voice dialogue scene according to the recognized specified words. Among them, the specified words can be stored by building a dictionary or word database. For example, the designated words corresponding to the closed-domain voice dialogue scene about the restaurant may include: I want to eat, I am hungry, etc., which indicate that I want to eat, or may be specific food such as hot pot, small noodles, or steak. Furthermore, the designated words corresponding to the closed-domain voice dialogue scene about travel may include: I want to go to Shenzhen Bay, I want to go to a football game, etc. to indicate travel intention.

Step S230: If not, it is determined that it is in an open-domain voice dialogue scenario, and when it is recognized that multiple feedback content can be used to respond to the received voice information, one feedback content is randomly selected for feedback.

Among them, as a way, if it is not a closed-domain voice dialogue scenario, when the voice information is received, the voice information is segmented to obtain multiple words, a card is displayed, and the card is displayed based on the multiple The content of the search term.

Step S240: If yes, recognize the voice information based on multiple slots corresponding to the target search requirement.

Step S250: If the target parameter corresponding to the target slot of the plurality of slots is successfully recognized from the voice information, a search instruction carrying the target parameter corresponding to each of the plurality of slots is generated, so The target parameter corresponding to the target slot is a parameter about the restaurant.

Step S260: Send the search instruction to the target restaurant information providing platform, and receive restaurant information returned by the target restaurant information providing platform in response to the search instruction.

In the embodiment of the present application, the voice assistant has multiple ways to obtain restaurant information. As a way, the target restaurant information providing platform is a client for providing restaurant information. In this way, the voice assistant can obtain restaurant information based on a data interface with the client for providing restaurant information. In this way, the voice assistant will send a search instruction generated based on the foregoing method to the client for providing restaurant information, and the search instruction is used to trigger the client for providing restaurant information to obtain according to specified rules Restaurant information; receiving the restaurant information returned by the client for providing restaurant information. As shown in Figure 7, in this way, the voice assistant will first perform the communication channel 95 and send the search command to the client for providing restaurant information. The client for providing restaurant information can first query whether there is a cache locally. The target restaurant information, if it is available, is directly returned to the voice assistant. If no cached target restaurant information is detected, a request for obtaining target restaurant information is sent to the server based on the communication channel 94, and then through the communication channel 94 and the communication channel 95 in turn Return to the voice assistant.

It is understandable that the client used to provide restaurant information here is a client that can be used to independently complete shopping order generation and order placement, similar to the Dianping client or Meituan client. As another way, a search instruction is sent to the server corresponding to the client for providing restaurant information, and the search instruction is used to trigger the server to obtain the target restaurant information according to a specified rule; and receive the information returned by the server Target restaurant information. For example, please refer to FIG. 7 again. In this manner, the voice assistant will directly send a search command to the server through the communication channel 93, and then the restaurant information will still be returned to the voice assistant through the communication channel 93. In yet another way, the voice assistant can forward the search instruction to the server through the client for providing restaurant information, and the server directly returns the restaurant information to the voice assistant instead of providing restaurant information. Of the client to convert, thereby increasing the information transmission rate. It is understandable that, for devices other than the electronic device where the voice assistant is located, the application program is distinguished by the port number occupied by the application program. For example, the browser client occupies port 80 and the voice assistant occupies port 8080. If the returned information points to port 8080, the electronic device will know that this information is returned to the voice assistant. In this way, the voice assistant adds the port number occupied by the voice assistant to the generated search command, and then when the server generates the returned information, the voice assistant’s port number is added to the returned information so that the information can be Send directly to the voice assistant.

Step S270: Display a card, and display the restaurant information in the card.

The information recommendation method provided by the present application realizes the slot required by configuring the search requirement, and then generates the subsequent search instruction based on the set slot, thereby improving the feedback for restaurant search Accuracy. In addition, in this embodiment, a closed-domain dialogue scenario and an open-domain dialogue scenario are configured, so that the voice assistant will only start to perform the voice analysis based on the multiple slots corresponding to the target search requirement after recognizing that it has entered the closed-domain dialogue scenario. Information is recognized, thereby improving the accuracy of recognition and reducing power consumption at the same time.

Referring to FIG. 8, an information recommendation method provided by an embodiment of the present application is applied to a voice assistant, and the method includes:

Step S310: Acquire voice information.

Step S320: Recognizing the voice information based on multiple slots corresponding to the target search requirement.

Step S330: If the target parameter corresponding to the target slot of the multiple slots is not successfully recognized from the voice information, the process ends.

Step S331: If the target parameter corresponding to the target slot of the multiple slots is successfully identified from the voice information, obtain the parameters corresponding to each of the multiple slots identified from the voice information quantity.

Step S332: Identify the quantity.

Step S340: If the number of parameters corresponding to the multiple slots is one, the identified parameters corresponding to the multiple slots are used as the target parameters of each slot.

Step S350: If there are multiple slots in the multiple slots, the number of corresponding parameters is multiple slots as the slot to be processed.

Step S360: Determine one parameter from the multiple parameters corresponding to the slot to be processed as the target parameter of the slot to be processed based on the target rule.

In this embodiment, the voice assistant may have multiple ways to determine the target parameter from multiple parameters.

As a way, the voice assistant can identify the positions in the voice information of multiple parameters corresponding to the slots to be processed; and use the parameters after the specified attribute words in the voice information as the target parameters. Optionally, the specified attribute words can be stored by establishing a dictionary or word database. The words that specify attributes can be words that represent actions. For example, eat, come, go, or play. For example, if the voice message recognized by the voice assistant is "I am in Chengdu, I am going to Chongqing to eat hot pot", then the parameters corresponding to this slot at the restaurant location include Chengdu and Chongqing, and furthermore, it is assigned to Chongqing as a designated attribute After the word "go", Chongqing is determined as the target parameter corresponding to the slot of the restaurant location.

As another way, the voice assistant may determine a parameter from multiple parameters corresponding to the slot to be processed as the target parameter of the slot to be processed based on at least one of the user's location information, historical portrait information, and user selection.

Step S370: Generate a search instruction carrying the target parameter corresponding to each of the multiple slots, and the target parameter corresponding to the target slot is a parameter about a restaurant.

Step S380: Send the search instruction to the target restaurant information providing platform, and receive restaurant information returned by the target restaurant information providing platform in response to the search instruction.

Step S390: Display a card, and display the restaurant information in the card.

The information recommendation method provided by the present application realizes the slot required by configuring the search requirement, and then generates the subsequent search instruction based on the set slot, thereby improving the feedback for restaurant search Accuracy. And in the case of identifying that there are multiple parameters corresponding to a slot, one parameter can be selected as the target parameter of the slot according to the target rule, so that it can still be more accurately identified in a more complex language environment The real intention of the user.

Please refer to FIG. 9, an information recommendation method provided by an embodiment of the present application is applied to a voice assistant, and the method includes:

Step S410: Acquire voice information.

Step S420: After receiving the voice information, identify the length of the received voice information;

Step S421: If the length is less than the specified length, search directly based on the received voice information, and display the searched content.

Step S430: If the length is not less than the specified length, recognize the voice information based on multiple slots corresponding to the target search requirement.

Step S440: If the target parameter corresponding to the target slot of the multiple slots is successfully recognized from the voice information, a search instruction carrying the target parameter corresponding to each of the multiple slots is generated, so The target parameter corresponding to the target slot is a parameter about the restaurant.

Step S450: Send the search instruction to the target restaurant information providing platform, and receive restaurant information returned by the target restaurant information providing platform in response to the search instruction.

Step S460: Display a card, and display the restaurant information in the card.

This application provides an information recommendation method, which realizes that after recognizing that the length of the voice message is not less than the specified length, the slot required by the search requirement is configured, and then the subsequent search is generated based on the set slot Instructions, which in turn improves the accuracy of feedback for restaurant searches.

Please refer to FIG. 10, an information recommendation apparatus 500 provided by an embodiment of the present application, the apparatus 500 includes:

The voice information receiving unit 510 is used to obtain voice information.

The voice recognition unit 520 is configured to recognize the voice information based on multiple slots corresponding to the target search requirement.

As a way, the voice recognition unit 520 is specifically configured to obtain the target search requirement locally after receiving the voice information; and recognize the voice information based on multiple slots corresponding to the obtained target search requirement. In this manner, as shown in FIG. 11, the device further includes a requirement obtaining unit 521, configured to obtain a target search request from the network, and receive an updated target search requirement obtained from the network; The target search demand of covers the local target search demand.

As a way, the voice recognition unit 520 is specifically configured to recognize the length of the received voice information; if the length is not less than a specified length, recognize the voice information based on multiple slots corresponding to the target search requirement. If the length is less than the specified length, perform semantic recognition on the received voice information. In this way, the card is displayed, and the content corresponding to the recognized semantics is displayed in the card.

The instruction generation unit 530 is configured to, if the target parameter corresponding to the target slot of the plurality of slots is successfully recognized from the voice information, generate a target parameter that carries the target parameter corresponding to each of the plurality of slots. In the search instruction, the target parameter corresponding to the target slot is a parameter about the restaurant.

The restaurant information acquiring unit 540 is configured to send the search instruction to the target restaurant information providing platform, and receive restaurant information returned by the target restaurant information providing platform in response to the search instruction.

The information display unit 550 is configured to display a card, and display the restaurant information in the card.

As a way, as shown in FIG. 12, the apparatus 500 further includes:

The scene processing unit 560 is configured to determine whether it is in a closed-domain voice dialogue scene about the restaurant based on the received voice. In this way,

If yes, the voice recognition unit 520 executes the multiple slots corresponding to the target search requirements to recognize the voice information; wherein, when in a closed-domain voice dialogue scene about the restaurant, after completing the voice information recognition If it is recognized that the slot lacks a target parameter, the voice assistant will issue a prompt message to prompt the user to enter the target parameter required by the slot lacking the target parameter.

If not, the voice recognition unit 520, after receiving the voice information, performs word segmentation on the voice information to obtain multiple words. The information display unit 550 is configured to display a card, and display the content after searching based on the multiple words in the card.

The scene processing unit 560 is also used for determining that it is in a closed-domain voice conversation scene about a restaurant in an open-domain voice dialogue scene when it is recognized that there are designated words related to dining in the received voice; In the scenario, the voice assistant randomly selects one feedback content for feedback when recognizing that there are multiple feedback content available for responding to the received voice information.

As a way, the instruction generating unit 530 is specifically configured to obtain the number of parameters corresponding to each of the multiple slots recognized from the voice information; if the number of parameters corresponding to the multiple slots is all One is to use the identified parameters corresponding to each of the multiple slots as the target parameters of each slot; generate a search instruction that carries the target parameters corresponding to each of the multiple slots.

In this manner, the instruction generating unit 530 is specifically further configured to, if there are multiple slots corresponding to the number of parameters in the multiple slots, set the number of corresponding parameters to multiple slots as Slot to be processed: Based on the target rule, a parameter is determined from multiple parameters corresponding to the slot to be processed as the target parameter of the slot to be processed.

Optionally, the instruction generating unit 530 is specifically configured to identify the positions in the voice information of multiple parameters corresponding to the slot to be processed; and use the parameters after the specified attribute word in the voice information as the target parameter . Wherein, the designated attribute words are words that characterize user actions.

Optionally, the instruction generation unit 530 is specifically configured to determine a parameter from the multiple parameters corresponding to the slot to be processed as the slot to be processed based on at least two conditions of the user's location information, historical portrait information, and user selection The target parameter.

Optionally, the instruction generation unit 530 is specifically configured to determine a parameter from multiple parameters corresponding to the slot to be processed as the target parameter of the slot to be processed based on the historical portrait information of the user.

Optionally, the instruction generation unit 530 is specifically configured to determine a parameter from multiple parameters corresponding to the slot to be processed as the target parameter of the slot to be processed based on the user's selection.

It should be noted that the device embodiment in this application and the foregoing method embodiment correspond to each other, and the specific principles in the device embodiment can be referred to the content in the foregoing method embodiment, which is not repeated here.

Hereinafter, an electronic device provided by the present application will be described with reference to FIG. 13.

Referring to FIG. 13, based on the above-mentioned information recommendation method and apparatus, an embodiment of the present application also provides another electronic device 200 that can execute the foregoing information recommendation method. The electronic device 200 includes one or more (only one is shown in the figure) a processor 102, a memory 104, and a network module 106 coupled to each other. Wherein, the memory 104 stores a program that can execute the content in the foregoing embodiment, and the processor 102 can execute the program stored in the memory 104. The processor 102 may include one or more processing cores. The processor 102 uses various interfaces and lines to connect various parts of the entire electronic device 200, and executes by running or executing instructions, programs, code sets, or instruction sets stored in the memory 104, and calling data stored in the memory 104. Various functions and processing data of the electronic device 200. Optionally, the processor 102 may use at least one of digital signal processing (Digital Signal Processing, DSP), Field-Programmable Gate Array (Field-Programmable Gate Array, FPGA), and Programmable Logic Array (Programmable Logic Array, PLA). A kind of hardware form to realize. The processor 102 may integrate one or a combination of a central processing unit (CPU), a graphics processing unit (GPU), a modem, and the like. Among them, the CPU mainly processes the operating system, user interface, and application programs; the GPU is used for rendering and drawing of display content; the modem is used for processing wireless communication. It can be understood that the above-mentioned modem may not be integrated into the processor 102, but may be implemented by a communication chip alone. The memory 104 may include random access memory (RAM) or read-only memory (Read-Only Memory). The memory 104 may be used to store instructions, programs, codes, code sets or instruction sets. The memory 104 may include a storage program area and a storage data area, where the storage program area may store instructions for implementing the operating system and instructions for implementing at least one function (such as touch function, sound playback function, image playback function, etc.) , Instructions for implementing the following method embodiments, etc. The data storage area can also store data (such as phone book, audio and video data, chat record data) created by the terminal 100 during use. The network module 106 is used to receive and send electromagnetic waves, realize the mutual conversion between electromagnetic waves and electrical signals, so as to communicate with a communication network or other devices, such as with audio playback devices. The network module 106 may include various existing circuit elements for performing these functions, for example, an antenna, a radio frequency transceiver, a digital signal processor, an encryption/decryption chip, a subscriber identity module (SIM) card, a memory, etc. . The network module 106 can communicate with various networks such as the Internet, an intranet, and a wireless network, or communicate with other devices through a wireless network. The aforementioned wireless network may include a cellular telephone network, a wireless local area network, or a metropolitan area network. For example, the network module 106 can exchange information with the base station. Please refer to FIG. 14, which shows a structural block diagram of a computer-readable storage medium provided by an embodiment of the present application. The computer-readable medium 1100 stores program code, and the program code can be invoked by a processor to execute the method described in the foregoing method embodiment. The computer-readable storage medium 1100 may be an electronic memory such as flash memory, EEPROM (Electrically Erasable Programmable Read Only Memory), EPROM, hard disk, or ROM. Optionally, the computer-readable storage medium 1100 includes a non-transitory computer-readable storage medium. The computer-readable storage medium 1100 has a storage space for executing the program code 810 of any method step in the foregoing method. These program codes can be read out from or written into one or more computer program products. The program code 1110 may be compressed in a suitable form, for example.

Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the application, not to limit them; although the application has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that: The technical solutions recorded in the foregoing embodiments are modified, or some of the technical features are equivalently replaced; these modifications or replacements do not drive the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

An information recommendation method, characterized in that it is applied to a voice assistant, and the method includes:

Get voice information;

Recognizing the voice information based on multiple slots corresponding to the target search requirement;

If the target parameter corresponding to the target slot of the multiple slots is successfully identified from the voice information, a search instruction carrying the target parameter corresponding to each of the multiple slots is generated, and the target slot The target parameter corresponding to the bit is the parameter about the restaurant;

Sending the search instruction to a target restaurant information providing platform, and receiving restaurant information returned by the target restaurant information providing platform in response to the search instruction;

A card is displayed, and the restaurant information is displayed in the card.
The method according to claim 1, wherein before the step of recognizing the voice information based on multiple slots corresponding to the target search requirement, the method further comprises:

Based on the received voice, determine whether it is in a closed-domain voice dialogue scene about the restaurant;

If yes, execute the multiple slots corresponding to the target search requirement to recognize the voice information; wherein, when in a closed-domain voice dialogue scene about the restaurant, after completing the voice information recognition, if the voice information is recognized If the slot lacks a target parameter, the voice assistant will issue a prompt message to prompt the user to enter the target parameter required by the slot lacking the target parameter.
The method of claim 2, wherein the method further comprises:

If not, when the voice information is received, the voice information is segmented to obtain multiple words, a card is displayed, and the content of the search based on the multiple words is displayed in the card.
The method according to claim 2 or 3, wherein the method further comprises:

In the open-domain voice dialogue scenario, when it is recognized that there are designated words about dining in the received voice, it is determined that it is in the closed-domain voice dialogue scenario about the restaurant; wherein, when in the open-domain dialogue scenario, the voice assistant is When it is recognized that multiple feedback content can be used to respond to the received voice information, one feedback content is randomly selected for feedback.
The method according to any one of claims 1 to 4, wherein the step of generating a search instruction carrying the target parameter corresponding to the slot comprises:

Acquiring the number of parameters corresponding to each of the multiple slots identified from the voice information;

If the number of parameters corresponding to the multiple slots is one, the identified parameters corresponding to the multiple slots are used as the target parameters of each slot;

Generate a search instruction carrying target parameters corresponding to each of the multiple slots.
The method according to claim 5, wherein the step of generating the search instruction carrying the parameter corresponding to the slot further comprises:

If the number of parameters corresponding to the slots is multiple in the multiple slots, the number of the corresponding parameters is multiple slots as the slot to be processed;

A parameter is determined from multiple parameters corresponding to the slot to be processed as the target parameter of the slot to be processed based on the target rule.
The method according to claim 6, wherein the step of determining a parameter from a plurality of parameters corresponding to the slot to be processed as the target parameter of the slot to be processed based on the target rule comprises:

Identifying positions in the voice information of multiple parameters corresponding to the slot to be processed;

The parameter after the designated attribute word in the voice information is taken as the target parameter.
The method according to claim 7, wherein the designated attribute words are words that characterize user actions.
The method according to claim 6, wherein the step of determining a parameter from a plurality of parameters corresponding to the slot to be processed as the target parameter of the slot to be processed based on the target rule comprises:

A parameter is determined from multiple parameters corresponding to the slot to be processed as a target parameter of the slot to be processed based on at least two conditions of the user's location information, historical portrait information and the user's selection.
The method according to claim 6, wherein the step of determining a parameter from a plurality of parameters corresponding to the slot to be processed as the target parameter of the slot to be processed based on the target rule comprises:

Based on the location information of the user, a parameter is determined from multiple parameters corresponding to the slot to be processed as the target parameter of the slot to be processed.
The method according to claim 6, wherein the step of determining a parameter from a plurality of parameters corresponding to the slot to be processed as the target parameter of the slot to be processed based on the target rule comprises:

Based on the user's historical portrait information, one parameter is determined from multiple parameters corresponding to the slot to be processed as the target parameter of the slot to be processed.
The method according to claim 6, wherein the step of determining a parameter from a plurality of parameters corresponding to the slot to be processed as the target parameter of the slot to be processed based on the target rule comprises:

Based on the user's selection, a parameter is determined from multiple parameters corresponding to the slot to be processed as the target parameter of the slot to be processed.
The method according to claim 1, wherein the step of recognizing the voice information based on multiple slots corresponding to the target search requirement comprises:

After receiving the voice information, obtain the target search requirements from the local;

The voice information is recognized based on the multiple slots corresponding to the acquired target search requirements.
The method of claim 13, wherein the method further comprises:

Receive updated target search requirements obtained from the Internet;

Make the updated target search demand cover the local target search demand.
The method according to claim 1, wherein the step of recognizing the voice information based on a plurality of slots corresponding to a target search requirement comprises:

Identifying the length of the received voice information;

If the length is not less than the specified length, the voice information is recognized based on multiple slots corresponding to the target search requirement.
The method according to claim 15, wherein the step of recognizing the voice information based on multiple slots corresponding to the target search requirement further comprises:

If the length is less than the specified length, perform semantic recognition on the received voice information;

A card is displayed, and content corresponding to the identified semantics is displayed in the card.
An information recommendation device, characterized in that the device includes:

The voice information receiving unit is used to obtain voice information;

The voice recognition unit is configured to recognize the voice information based on multiple slots corresponding to the target search requirement;

The instruction generating unit is configured to generate a search carrying the target parameter corresponding to each of the multiple slots if the target parameter corresponding to the target slot of the multiple slots is successfully recognized from the voice information Instruction, the target parameter corresponding to the target slot is a parameter about the restaurant;

A restaurant information acquisition unit, configured to send the search instruction to a target restaurant information providing platform, and receive restaurant information returned by the target restaurant information providing platform in response to the search instruction;

The information display unit is used to display a card and display the restaurant information in the card.
The device according to claim 17, wherein the device further comprises:

The scene processing unit is used to determine whether it is in a closed-domain voice dialogue scene about the restaurant based on the received voice;

If yes, the voice recognition unit executes the multiple slots corresponding to the target search requirement to recognize the voice information; wherein, when in a closed-domain voice dialogue scene about the restaurant, after completing the voice information recognition, If it is recognized that the slot lacks a target parameter, the voice assistant will issue a prompt message to prompt the user to enter the target parameter required by the slot lacking the target parameter.
An electronic device, characterized in that it includes one or more processors and a memory;

One or more programs are stored in the memory and configured to be executed by the one or more processors to implement the method of any one of claims 1-16.
A computer-readable storage medium, wherein a program code is stored in the computer-readable storage medium, wherein the method according to any one of claims 1-16 is executed when the program code is run by a processor.