Data processing method, device and equipment
By setting multiple information recognition modules in the data processing device and determining weights based on the request information, calculating the matching value of the entity information and the module, the problem that the vehicle-mounted voice assistant is prone to errors during text recognition is solved, and the naturalness and accuracy of human-computer interaction is improved, and the security risks are reduced.
Patent Information
- Application Number
- CN202411988433.2
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-12-31
- Publication Date
- 2025-05-16
AI Technical Summary
Existing vehicle voice assistants are prone to errors when recognizing text, resulting in poor user experience, and during the recognition process, users need to focus on the screen, affecting safe driving.
By setting a plurality of information identification modules in the data processing device, determining the weight of each module based on the request information, calculating the matching value of the entity information and the module, and then determining the target entity information, improving the naturalness and accuracy of the interaction.
It improves the smoothness, nature and accuracy of human-computer interaction, reduces users' dependence on the screen during driving, and reduces safety risks.
Smart Images
Figure CN120010658A_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the field of data processing, and in particular to a data processing method, device and equipment. Background Art
[0002] With the development of technology, the products people use in daily life are becoming more and more intelligent. In order to improve people's driving experience, in-car voice assistants have come into being. In-car voice assistants provide users with a more natural way of communication. In-car voice assistants can provide users with a natural language interaction method, and users can directly send instructions to the voice assistant to improve convenience.
[0003] For example, if a user wants to share a string of numbers (such as a phone number) on the car screen with a friend, without the help of a voice assistant, the user may use the following two methods to achieve this: 1. Remember the number → Enter Messages → Send a new message to a friend → Enter the number from memory → Send. 2. Select the number → Copy → Go to Messages → Send a new message to a friend → Paste → Send. However, these two methods require users to click, switch applications, scroll, and enter multiple times.
[0004] After the birth of voice assistants, users can use voice assistants to: read the number to the in-car voice assistant and tell the in-car voice assistant to send a new message to a friend. This implementation method is complicated, unnatural and not very intelligent. Although this implementation method does not require a touch screen, it requires the user to focus on the screen to read the number or requires the user to remember the number in advance. Focusing on the screen will have a great negative impact on safe driving, resulting in a poor user experience. In addition, voice assistants are prone to errors when performing text recognition, which further leads to a poor user experience. Summary of the invention
[0005] In view of this, embodiments of the present application provide a data processing method, apparatus, and device to improve the naturalness and accuracy of human-computer interaction.
[0006] A data processing method, the data processing method is applicable to a data processing device, the data processing device includes a plurality of information identification modules; the method includes:
[0007] Get request information;
[0008] Based on the request information, determining a weight corresponding to each information identification module in the data processing device;
[0009] Based on the request information, obtaining an entity information set;
[0010] Obtaining a matching value between each entity information and each information identification module in the entity information set;
[0011] Based on the matching value between each entity information in the entity information set and each information identification module and the weight corresponding to each information identification module, target entity information is determined in the entity information set.
[0012] In a possible implementation manner of the first aspect, the request information includes voice information.
[0013] In a possible implementation manner of the first aspect, determining, based on the request information, a weight corresponding to each information identification module in the data processing device includes:
[0014] Convert request information into a request embedding vector;
[0015] Based on the request embedding vector, a weight corresponding to each information identification module in the data processing device is obtained.
[0016] In a possible implementation of the first aspect, obtaining the weight corresponding to each information identification module in the data processing device based on the request embedding vector includes: processing the request embedding vector based on a multilayer perceptron and a normalized exponential function to obtain the weight corresponding to each information identification module in the data processing device.
[0017] In a possible implementation manner of the first aspect, the sum of the weights corresponding to all information identification modules in the data processing device is 1.
[0018] In a possible implementation manner of the first aspect, acquiring the entity information set based on the request information includes:
[0019] Parsing the address information of the entity information set from the request information;
[0020] Based on the address information, the entity information set is obtained.
[0021] In a possible implementation manner of the first aspect, the entity information in the entity information set includes text information.
[0022] In a possible implementation manner of the first aspect, determining the target entity information in the entity information set based on a matching value between each entity information in the entity information set and each information identification module and a weight corresponding to each information identification module includes:
[0023] Based on the matching value between each entity information and each identification module and the weight corresponding to each information identification module, a matching value between each entity information in the entity information set and the request information is obtained according to a weighted sum algorithm;
[0024] The entity information in the entity information set whose matching value with the request information satisfies a preset condition is used as the target entity information.
[0025] In a possible implementation manner of the first aspect, a matching value between the entity information m and the request information satisfies Dm:
[0026]
[0027] Among them, ki represents the weight of information identification module i, and Zim represents the matching value between entity information m and information identification module i.
[0028] In a possible implementation manner of the first aspect, the multiple information identification modules include: at least one of: a category module, a location module, and a text module.
[0029] In a second aspect, an embodiment of the present application provides a data processing device, including multiple information identification modules, and the data processing device further includes: an acquisition module, a weight calculation module, and a processing module. Among them, the acquisition module is used to obtain request information; the weight calculation module is used to determine the weight corresponding to each information identification module in the data processing device based on the request information; the processing module is used to obtain an entity information set based on the request information; the processing module is also used to determine the target entity information in the entity information set based on the request information and the weight corresponding to each information identification module.
[0030] In a third aspect, an embodiment of the present application provides a vehicle, comprising the device provided in the second aspect.
[0031] In a fourth aspect, an embodiment of the present application provides an electronic device comprising a memory for storing computer program instructions and a processor for executing the program instructions, wherein when the computer program instructions are executed by the processor, the electronic device executes the method provided in the first aspect.
[0032] In a fifth aspect, an embodiment of the present application provides a computer-readable storage medium, wherein the computer-readable storage medium includes a stored program, wherein when the program is running, the device where the computer-readable storage medium is located is controlled to execute the method provided in the first aspect.
[0033] In a sixth aspect, an embodiment of the present application provides a computer program product, which includes executable instructions. When the executable instructions are executed on a computer, the computer executes the method provided in the first aspect.
[0034] The data processing method provided in the embodiment of the present application assigns a weight to each information identification module, calculates the matching value between the entity information and the information identification module, and then determines the matching value between the entity information and the request information based on the weight and the matching value between the entity information and the information identification module, thereby determining entity information with higher accuracy and more in line with user expectations in the entity information set. Therefore, the embodiment of the present application can improve the smoothness, naturalness and accuracy of human-computer interaction. BRIEF DESCRIPTION OF THE DRAWINGS
[0035] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.
[0036] Figure 1 A flowchart of a data processing method provided in an embodiment of the present application;
[0037] Figure 2 A flowchart for determining a weight corresponding to each information identification module in a data processing device provided in an embodiment of the present application;
[0038] Figure 3 A flowchart of obtaining an entity information set based on request information provided in an embodiment of the present application;
[0039] Figure 4 A flowchart for determining target entity information in an entity information set based on request information and a weight corresponding to each information identification module provided in an embodiment of the present application;
[0040] Figure 5 A schematic diagram of a data processing device provided in an embodiment of the present application;
[0041] Figure 6 A schematic diagram of a data processing device provided in an embodiment of the present application;
[0042] Figure 7 A schematic diagram of an electronic device provided in an embodiment of the present application.
[0043] Description of symbols
[0044] 100. Data processing device; 101. Acquisition module; 102. Weight calculation module; 103. Processing module; 104. Category module; 105. Location module; 106. Text module; 200. Electronic device; 201. Processor; 202. Memory; 203. Communication unit. DETAILED DESCRIPTION
[0045] In order to better understand the technical solution of the present application, the embodiments of the present application are described in detail below with reference to the accompanying drawings.
[0046] It should be clear that the described embodiments are only part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in the field without creative work are within the scope of protection of the present application.
[0047] The terms used in the embodiments of the present application are only for the purpose of describing specific embodiments, and are not intended to limit the present application. The singular forms "a", "said" and "the" used in the embodiments of the present application and the appended claims are also intended to include plural forms, unless the context clearly indicates other meanings.
[0048] It should be understood that the term "and / or" used in this article is only a description of the association relationship of associated objects, indicating that there can be three relationships. For example, A and / or B can represent: A exists alone, A and B exist at the same time, and B exists alone. In addition, the character " / " in this article generally indicates that the associated objects before and after are in an "or" relationship.
[0049] The embodiment of the present application provides a data processing method, which is applicable to a data processing device, and the data processing device includes a plurality of information recognition modules. The information recognition modules are used to perform information feature recognition.
[0050] In a possible implementation, the multiple information identification modules include: at least one of a category module, a location module, and a text module.
[0051] Among them, the category module focuses on the category information of the entity information, such as phone number, email address, address, etc. The location module focuses on the location information of the entity information, such as its coordinates on the screen. The text module focuses on the text information of the entity information, such as the words it contains.
[0052] like Figure 1 As shown, a data processing method provided in an embodiment of the present application includes:
[0053] S100: Obtain request information.
[0054] The request information is sent by the user. In one possible implementation, the request information includes a wake-up request and an execution request. The wake-up request is used to wake up the data processing device, such as waking up the voice assistant, and can be awakened by an operation instruction: for example, operating the center console in a preset manner, such as long pressing or double-clicking. The voice assistant can also be awakened by natural language, such as by calling the voice assistant nickname. The execution request contains instruction information, which is used to represent the instructions issued by the user, that is, the tasks that need to be performed by the data processing device, such as sending the phone number of XX (contact a) to YY (contact b or a designated address, such as a designated phone number or a designated email address, etc.), or sharing yy at xx (yy refers to information, such as a phone number, email, home address, document information, etc.) with zz (for example, a target contact or target address, etc.).
[0055] In a possible implementation, when the user issues a request by operating the central console, the data processing device directly receives the request. When the user issues a request in natural language, the data processing device acquires the request by capturing or sensing natural speech.
[0056] In a possible implementation manner, the request information includes voice information.
[0057] Voice information is information in the form of voice, such as natural language uttered by a user or other voice device. For example, the phone number of XX (contact a) is sent to YY (contact b or a specified address, such as a specified phone number or a specified email address).
[0058] In this implementation, the request information includes voice information, which means that a voice assistant can be used to issue commands during human-computer interaction, thereby solving the problem of using both hands, increasing the safety factor during driving, and improving convenience of use.
[0059] S200: Determine the weight corresponding to each information identification module in the data processing device based on the request information.
[0060] The information recognition module is used to recognize information features. Different information modules recognize different information features. For example, the text module is used to recognize text in information, and the category module is used to recognize the category of information.
[0061] In step S200, based on the request content of the request information, the data processing device will assign weights to each information identification module, so that the parsing process of the request information can be more accurate, the human-computer interaction can be smoother, and the response based on the contextual meaning can be the same as the user's expectations. For example, the request information is: call the top phone number. Since the request information contains the location keyword "top" and the category keyword "phone number", the data processing device will assign higher weights to the location module and the category module.
[0062] like Figure 2 As shown, in a possible implementation, based on the request information, determining the weight corresponding to each information identification module in the data processing device includes:
[0063] S210: Convert the request information into a request embedding vector.
[0064] After receiving the request information, the data processing device converts the request information into a request embedding vector through the first embedder in the data processing device. The request embedding vector is the basis for determining the weight of each information recognition module and the basis for determining the target text information.
[0065] S220. Obtain a weight corresponding to each information identification module in the data processing device based on the request embedding vector.
[0066] After the data processing device obtains the request embedding vector, the request embedding vector is passed through the weight calculation module to obtain the weight corresponding to each information identification module. In one possible implementation, the weight calculation module includes a multi-layer perceptron (MLP) module and a normalized exponential function (softmax function). The MLP calculates the weight of each information identification module based on the request embedding vector, and the softmax function normalizes these weights so that their sum is 1. Therefore, the sum of the weights corresponding to all information identification modules in the data processing device is 1. For example, the data processing device includes n information identification modules, and the weights of the n information identification modules are k1, k2, k3, ..., kn, respectively, then:
[0067] In step S200, the weight calculation module assigns different weights to each information identification module according to different request information. For example, for the request information "call the top phone number", the weight calculation module will assign higher weights to the location module and the category module because they contain the location keyword "top" and the type keyword "phone number". For the request information "call a certain colleague in Liuzhou, Guangxi", the weight calculation module may assign higher weights to the text module and the category module because they contain the text keywords "Liuzhou, Guangxi" and "colleague" and the category keyword "phone".
[0068] S300: Acquire an entity information set based on the request information.
[0069] After receiving the request information, the data processing device will obtain an entity information set according to the request information. The entity information set refers to a set containing target entity information. In a possible implementation, the entity information in the entity information set includes text information.
[0070] like Figure 3 As shown, in a possible implementation, based on the request information, obtaining the entity information set includes:
[0071] S310: Parse the address information of the entity information set from the request information.
[0072] The entity information set includes target entity information, and the target entity information is the entity information pointed to by the request. The data processing device can obtain the address information of the entity information set based on the request information. For example, the data processing device can obtain the address information of the entity information set as screen display information based on "calling the top phone number". For example, the data processing device can obtain the address information of the entity information set as the address book based on "calling a certain colleague in Liuzhou, Guangxi". The implementation of this process can be achieved by relying on the built-in data processing model of the data processing device. The data processing model is based on the result of multi-sample training, and can parse the address information of the target entity information from the relevant keywords in the request information. For example, it can parse the location of the target entity information as the display screen based on the keyword "top" in the request information, and the display screen is the address information of the entity information set; it can parse the location of the target entity information as the address book based on the pipe detection "call" in the request information, that is, the address book is the address information of the entity information set.
[0073] S320: Obtain an entity information set based on the address information.
[0074] For example, for the request information "call the top phone number", the basis of the information set is the content displayed on the current display screen, and the data processing device obtains the entity information set based on the display information of the current display screen. For example, the screen image is obtained by screenshot, and then the text is extracted from the screen image. For example, the data detector can identify and extract the text information in the screen image through accessibility, OCR, image recognition and other means. Or the screen display information is directly obtained through the processor of the central console to obtain the entity information set.
[0075] For example, for the request information "call a certain colleague in Liuzhou, Guangxi", the data processing device accesses the address book according to the request information, and all the address book information in the address book is an entity information set. Among them, a piece of communication information includes the contact name, phone number, address, email address, etc.
[0076] In a possible implementation manner, the entity information in the entity information set includes image information and / or video information.
[0077] It should be noted that there is no particular order in which step S200 and step S300 are executed. Figure 1 As shown, step S200 is performed first, and then step S300 is performed. Step S300 may also be performed first and then step S200. Steps S200 and S300 may also be performed simultaneously.
[0078] S400: Obtain a matching value between each entity information in the entity information set and each information identification module.
[0079] In step S400, the data processing device calculates the matching value between each entity information and each information identification module in the entity information set according to the request information and the weight corresponding to each identification module. The matching value between the entity information and the information identification module can represent the matching degree between the entity information and the information identification module. The larger the matching value, the higher the matching degree. The matching value can be expressed by a numerical value (such as a score value) or a percentage.
[0080] In step S400, the request embedding vector is input into the embedder of each information identification module. The embedder of each information identification module applies a soft attention mechanism to focus on the request part related to the module and obtain an embedding vector. Each information identification module calculates the score Zim of each entity information based on the characteristics of the entity information and the embedding vector in its embedder, where i represents the information identification module and m represents the entity information.
[0081] For example, the category module compares the entity category embedding with the request embedding (i.e., the embedding vector in the category module) and calculates a score based on the degree of match. The location module compares the entity's bounding box features with the request embedding (i.e., the embedding vector in the location module) and calculates a score based on the positional relationship. The text module compares the entity text with the request embedding (i.e., the embedding vector in the text module) and calculates a match value based on the degree of text match.
[0082] S500: Determine target entity information in the entity information set based on a matching value between each entity information in the entity information set and each information identification module and a weight corresponding to each information identification module.
[0083] In step S500, the matching value between each entity information and the request information can be obtained through the matching value between each entity information and each information identification module and the weight corresponding to each information identification module, and then the entity information with the matching value that meets the preset conditions is taken as the target entity information. For example, the entity information with the largest matching value with the request information is taken as the target entity information.
[0084] like Figure 4 As shown, in a possible implementation, based on the matching value between each entity information in the entity information set and each information identification module and the weight corresponding to each information identification module, determining the target entity information in the entity information set includes:
[0085] S510 : Based on the matching value between each entity information in the entity information set and each information identification module and the weight corresponding to each information identification module, obtain the matching value between each entity information in the entity information set and the request information according to a weighted sum algorithm.
[0086] The matching value of the same entity information and each information recognition module is multiplied by the corresponding weight, and then the sum is obtained to obtain the matching value of the entity information and the request information. That is, the matching value Dm of the entity information m and the request information satisfies:
[0087]
[0088] Among them, ki represents the weight of information identification module i, * represents the multiplication sign, Zim represents the matching value between entity information m and information identification module i; n is the number of information identification modules, and i, n, and m are positive integers respectively.
[0089] Traverse all entity information sets and obtain the matching value between each entity information and the requested information.
[0090] For example, there are first, second and third entity information on the screen, wherein the first entity information is located at the top, the second entity information is located in the middle, and the third entity information is located at the bottom. The data processing device includes a text module, a location module and a category module. The request information "send the top phone number to A" is used as an example for detailed description.
[0091] Based on the request information, it is determined that the weight of the text module is k1, the weight of the position module is k2, and the weight of the category module is k3.
[0092] The text module will calculate the matching value Z11 between the first text information and the text module based on the characteristics of the first entity information and the embedding vector in its embedder. The text module will calculate the matching value Z12 between the second text information and the text module based on the characteristics of the second entity information and the embedding vector in its embedder. The text module will calculate the matching value Z13 between the second text information and the text module based on the characteristics of the third entity information and the embedding vector in its embedder.
[0093] The position module will calculate the matching value Z21 between the first text information and the position module based on the characteristics of the first entity information and the embedding vector in its embedder. The position module will calculate the matching value Z22 between the second text information and the position module based on the characteristics of the second entity information and the embedding vector in its embedder. The position module will calculate the matching value Z23 between the second text information and the position module based on the characteristics of the third entity information and the embedding vector in its embedder.
[0094] The category module will calculate the matching value Z31 between the first text information and the category module based on the characteristics of the first entity information and the embedding vector in its embedder. The category module will calculate the matching value Z32 between the second text information and the category module based on the characteristics of the second entity information and the embedding vector in its embedder. The category module will calculate the matching value Z33 between the second text information and the category module based on the characteristics of the third entity information and the embedding vector in its embedder.
[0095] Then there are:
[0096] The matching value D1 between the first entity information and the request information is D1=K1*Z11+K2*Z21+K3*Z31.
[0097] The matching value between the second entity information and the request information is D2=K1*Z12+K2*Z22+K3*Z32.
[0098] The matching value between the second entity information and the request information is D3=K1*Z13+K2*Z23+K3*Z33.
[0099] S520: Entity information in the entity information set whose matching value with the request information meets a preset condition is used as target entity information.
[0100] The preset condition is set based on the actual application scenario. In a possible implementation, the preset condition includes: the entity information with the highest matching value with the request information, that is, the entity information with the highest matching value with the request information in the entity information set is used as the target entity information.
[0101] For example, a maximum value is determined among D1, D2, and D3, for example, D1 is the maximum value, and the first entity information is the target entity information.
[0102] In an embodiment of the present application, after step S500, the following steps are further included:
[0103] S600, based on the request information, output processing is performed on the target text information. The output processing includes sending, calling, displaying, opening an application, copying, pasting, modifying, deleting, etc. The output processing method is determined based on the request information. For example, for the request information "send the top phone to A", after determining the target entity information, the output processing is to send the target entity information to A.
[0104] In summary, the embodiment of the present application determines the weight of each information identification module and the matching value between the entity information and the information identification module, and uses the weight value and the matching value between the entity information and the information identification module to obtain the matching value between the entity information and the request information, and then determines the target entity information based on the matching value between the entity information and the request information. In this way, the target entity information has a high degree of matching with the content requested in the user request information, thereby making human-computer interaction more natural, concise and accurate.
[0105] like Figure 5 As shown, an embodiment of the present application also provides a data processing device 100, comprising a plurality of information identification modules. The data processing device 100 further comprises: an acquisition module 101, a weight calculation module 102, and a processing module 103. The acquisition module 101 is used to obtain request information. The weight calculation module 102 is used to determine the weight corresponding to each information identification module in the data processing device based on the request information. The processing module 103 obtains the matching value between each entity information in the entity information set and each information identification module; the processing module 103 is also used to determine the target entity information in the entity information set based on the matching value between each entity information in the entity information set and each information identification module and the weight corresponding to each information identification module. The processing module 103 comprises a data processing model.
[0106] In a possible implementation, the multiple information identification modules include: at least one of a category module 104, a position module 105, and a text module 106. The category module 104 may be the category module 104 provided in any of the aforementioned embodiments. The position module 105 may be the position module 105 provided in any of the aforementioned embodiments. The text module 106 may be the text module 106 provided in any of the aforementioned embodiments. Figure 6 As shown, the multiple information identification modules include: a category module 104 , a location module 105 and a text module 106 .
[0107] In a possible implementation manner, the request information includes voice information.
[0108] In a possible implementation, the weight calculation module 102 determines the weight corresponding to each information identification module in the data processing device based on the request information, including: the weight calculation module 102 converts the request information into a request embedding vector. The weight calculation module 102 obtains the weight corresponding to each information identification module in the data processing device based on the request embedding vector.
[0109] In a possible implementation, the sum of the weights corresponding to all information identification modules in the data processing device is 1.
[0110] In a possible implementation, the processing module 103 obtains the entity information set based on the request information, including:
[0111] The processing module 103 parses the request information to obtain the address information of the entity information set;
[0112] The processing module 103 obtains an entity information set based on the address information.
[0113] The specific processing process of the processing module 103 in this implementation can refer to the embodiment of the aforementioned method, which will not be described in detail here.
[0114] In a possible implementation manner, the entity information in the entity information set includes text information.
[0115] In one possible implementation, the processing module 103 determines the target entity information in the entity information set based on the request information and the weight corresponding to each information identification module: the processing module 103 obtains the matching value between each entity information in the entity information set and each information identification module; the processing module 103 obtains the matching value between each entity information in the entity information set and the request information based on the matching value between each entity information in the entity information set and each information identification module and each information identification module; the processing module 103 takes the entity information in the entity information set whose matching value with the request information meets the preset conditions as the target entity information.
[0116] In this implementation, the specific processing process of the processing module 103 can refer to the embodiment of the aforementioned method, which will not be described in detail here.
[0117] An embodiment of the present application also provides a vehicle, comprising the data processing device provided in the aforementioned embodiment.
[0118] A vehicle using the data processing device provided in the embodiment of the present application can better use the voice assistant, making human-computer interaction more natural and accurate.
[0119] An embodiment of the present application also provides an electronic device, comprising a memory for storing computer program instructions and a processor for executing the program instructions, wherein when the computer program instructions are executed by the processor, the electronic device executes the data processing method provided by the aforementioned embodiment.
[0120] An electronic device that adopts the data processing device provided in the embodiment of the present application can better use the voice assistant, making human-computer interaction more natural and accurate.
[0121] like Figure 7 As shown, in a possible implementation, the electronic device 200 includes: a processor 201, a memory 202 and a communication unit 203. These components communicate through one or more buses. Those skilled in the art can understand that the structure of the electronic device shown in the figure does not constitute a limitation on the embodiment of the present invention. It can be a bus structure or a star structure, and can also include more or less components than shown in the figure, or combine some components, or arrange the components differently.
[0122] The communication unit 203 is used to establish a communication channel so that the electronic device can communicate with other devices, receive user data sent by other devices or send user data to other devices.
[0123] The processor 201 is the control center of the electronic device. It uses various interfaces and lines to connect various parts of the entire electronic device. It runs or executes software programs, instructions, and / or modules stored in the memory 202, and calls data stored in the memory to perform various functions of the electronic device and / or process data. The processor can be composed of an integrated circuit (IC), for example, it can be composed of a single packaged IC, or it can be composed of multiple packaged ICs with the same or different functions. For example, the processor 201 can only include a central processing unit (CPU). In an embodiment of the present invention, the CPU can be a single computing core or multiple computing cores.
[0124] The memory 202 is used to store the execution instructions of the processor 201. The memory 202 can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, magnetic disk or optical disk.
[0125] When the execution instructions in the memory 202 are executed by the processor 201, the electronic device 200 can execute Figure 1 Some or all of the steps in the illustrated embodiments.
[0126] The embodiment of the present application also provides a computer-readable storage medium, the computer-readable storage medium includes a stored program, wherein when the program is running, the device where the computer-readable storage medium is located is controlled to execute the method provided in any of the above embodiments. The storage medium can be a magnetic disk, an optical disk, a read-only memory (ROM) or a random access memory (RAM), etc.
[0127] An embodiment of the present application also provides a computer program product, which includes executable instructions. When the executable instructions are executed on a computer, the computer executes the data processing method provided by any of the aforementioned embodiments.
[0128] Those skilled in the art can clearly understand that the technology in the embodiments of the present invention can be implemented by means of software plus a necessary general hardware platform. Based on this understanding, the technical solution in the embodiments of the present invention is essentially or the part that contributes to the prior art can be embodied in the form of a software product, which can be stored in a storage medium such as ROM / RAM, a disk, an optical disk, etc., and includes a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment of the present invention or some parts of the embodiments.
[0129] In this specification, the same or similar parts between the various embodiments can be referred to each other. In particular, for the device embodiment and the terminal embodiment, since they are basically similar to the method embodiment, the description is relatively simple, and the relevant parts can be referred to the description in the method embodiment.
Claims
1. A data processing method, characterized in that: The data processing method is applicable to a data processing device, wherein the data processing device includes a plurality of information identification modules; the method includes: Get request information; Based on the request information, determining a weight corresponding to each information identification module in the data processing device; Based on the request information, obtaining an entity information set; Obtaining a matching value between each entity information and each information identification module in the entity information set; Based on the matching value between each entity information in the entity information set and each information identification module and the weight corresponding to each information identification module, target entity information is determined in the entity information set.
2. The method according to claim 1, characterized in that The request information includes voice information.
3. The method according to claim 1, characterized in that The determining, based on the request information, a weight corresponding to each information identification module in the data processing device comprises: Convert request information into a request embedding vector; Based on the request embedding vector, a weight corresponding to each information identification module in the data processing device is obtained.
4. The method according to claim 3, characterized in that The obtaining, based on the request embedding vector, a weight corresponding to each information identification module in the data processing device comprises: Based on a multi-layer perceptron and a normalized exponential function, the request embedding vector is processed to obtain a weight corresponding to each information identification module in the data processing device.
5. The method according to claim 1, characterized in that The sum of the weights corresponding to all the information identification modules in the data processing device is 1.
6. The method according to claim 1, characterized in that The acquiring of the entity information set based on the request information includes: Parsing the address information of the entity information set from the request information; Based on the address information, the entity information set is obtained.
7. The method according to claim 1, characterized in that The entity information in the entity information set includes text information.
8. The method according to claim 1, characterized in that The determining the target entity information in the entity information set based on the matching value between each entity information in the entity information set and each information identification module and the weight corresponding to each information identification module comprises: Based on the matching value between each entity information and each identification module and the weight corresponding to each information identification module, a matching value between each entity information in the entity information set and the request information is obtained according to a weighted sum algorithm; The entity information in the entity information set whose matching value with the request information satisfies a preset condition is used as the target entity information.
9. The method according to claim 8, characterized in that The matching value between the entity information m and the request information satisfies Dm: Among them, ki represents the weight of information identification module i, and Zim represents the matching value between entity information m and information identification module i.
10. The method according to claim 1, characterized in that The plurality of information identification modules include at least one of a category module, a location module, and a text module.
11. A data processing device, characterized in that: The data processing device comprises a plurality of information identification modules, and the data processing device further comprises: The acquisition module is used to obtain request information; A weight calculation module, the weight calculation module is used to determine the weight corresponding to each information identification module in the data processing device based on the request information; The processing module is used to obtain an entity information set based on the request information; The processing module is also used to obtain a matching value between each entity information in the entity information set and each information identification module; The processing module is further configured to determine target entity information in the entity information set based on a matching value between each entity information in the entity information set and each information identification module and a weight corresponding to each information identification module.
12. A vehicle, characterized in that: The vehicle comprises the device of claim 11.
13. An electronic device, characterized in that: The electronic device comprises a memory for storing computer program instructions and a processor for executing the program instructions, wherein when the computer program instructions are executed by the processor, the electronic device executes the method according to any one of claims 1 to 10.
14. A computer-readable storage medium, characterized in that: The computer-readable storage medium includes a stored program, wherein when the program is executed, the device where the computer-readable storage medium is located is controlled to execute the method according to any one of claims 1 to 10.
15. A computer program product, characterized in that The computer program product comprises executable instructions, and when the executable instructions are executed on a computer, the computer is caused to perform the method according to any one of claims 1 to 10.