Intelligent response method and apparatus
By analyzing the target consultation question using natural language processing and bidirectional long short-term memory networks, the problem of repetitive responses in intelligent response schemes was solved, resulting in more accurate and intelligent responses and improved user experience.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- CHINA TELECOM CORP LTD
- Filing Date
- 2024-03-25
- Publication Date
- 2026-07-21
AI Technical Summary
Existing intelligent response solutions provide independent responses in different scenarios during multi-turn dialogues, leading to repetitive responses and impacting user experience.
By analyzing the target consultation question and historical intent using natural language processing algorithms and bidirectional long short-term memory networks, it is determined whether there is an option selection intent in the target consultation intent. If so, the corresponding option is determined and the response information is fed back; otherwise, the response information is fed back directly.
It improves the accuracy of responses, reduces the repetitive response rate, optimizes the user experience, and reduces the complexity of user operations.
Smart Images
Figure CN118170888B_ABST
Abstract
Description
Technical Field
[0001] This application relates to the field of artificial intelligence technology, and more specifically, to an intelligent response method and apparatus. Background Technology
[0002] With the continuous development of human-computer interaction, intelligent response solutions have gradually acquired the ability to conduct multi-turn interactions based on topics. Traditional scene interaction can understand the intent of the robot based on communication with it. When the understood intent matches the triggering conditions of the scene, it enters the scene for more accurate and in-depth interactive communication.
[0003] However, in the long-term use of existing intelligent response solutions, taking the intelligent response scenario of mobile phone plans as an example, the same menu option appears in the plan scenario and the plan tier scenario. Users need to make the same selection repeatedly. In the above scenarios, the semantic training rules are presented as independent FAQs or independent scenarios, lacking a global understanding of user intent, which affects the user experience.
[0004] There is currently no effective solution to the above problems. Summary of the Invention
[0005] This application provides an intelligent response method and apparatus to at least solve the technical problem that independent responses in different scenarios in related intelligent response schemes can easily lead to repeated responses in multi-round dialogues, affecting user experience.
[0006] According to one aspect of the embodiments of this application, an intelligent response method is provided, comprising: obtaining a target consultation question for the current round and obtaining a historical intent list, wherein the historical intent list includes: historical consultation intents corresponding to historical consultation questions in each round prior to the current round in a multi-round dialogue; analyzing the target consultation question and each historical consultation intent using a natural language processing algorithm to obtain the target consultation intent corresponding to the target consultation question, and determining first response information corresponding to the target consultation intent; if the first response information includes options, analyzing each historical consultation intent and the target consultation intent using a bidirectional long short-term memory network to determine whether there is an option selection intent in the target consultation intent; if there is no option selection intent in the target consultation intent, feeding back the first response information; if there is an option selection intent in the target consultation intent, determining second response information based on the target option corresponding to the option selection intent, and feeding back the second response information.
[0007] Optionally, natural language processing algorithms are used to analyze the target consultation question and each historical consultation intent to obtain the target consultation intent corresponding to the target consultation question, including: using semantic analysis algorithms to analyze the target consultation question to obtain the target semantic information corresponding to the target consultation question; and using a bidirectional long short-term memory network to analyze each historical consultation intent and the target semantic information to obtain the target consultation intent corresponding to the target consultation question.
[0008] Optionally, determining the first response information corresponding to the target consultation intent includes: determining the first consultation scenario corresponding to the target consultation intent; and determining the first response information corresponding to the first consultation scenario from a preset mapping relationship table, wherein the mapping relationship table stores the mapping relationship between multiple consultation scenarios and multiple response information.
[0009] Optionally, after determining the first response information corresponding to the target consultation intent, the method further includes: if the first response information does not include options, feeding back the first response information.
[0010] Optionally, determining the second response information based on the target option corresponding to the option selection intent includes: determining the target option corresponding to the option selection intent from the options in the first response information, and determining the second consultation scenario corresponding to the target option; and determining the second response information corresponding to the second consultation scenario from the mapping relationship table.
[0011] Optionally, the second response information is fed back, including simultaneously feeding back the second response information and the other options in the first response information, excluding the target option.
[0012] Optionally, after providing feedback on the response, the method may also include adding the target consultation intent to a list of historical intents.
[0013] According to another aspect of the embodiments of this application, an intelligent response device is also provided, comprising: an acquisition module, configured to acquire the target consultation question of the current round and acquire a historical intent list, wherein the historical intent list includes: historical consultation intents corresponding to historical consultation questions of each round before the current round in a multi-round dialogue; a first analysis module, configured to analyze the target consultation question and each historical consultation intent using a natural language processing algorithm to obtain the target consultation intent corresponding to the target consultation question and determine the first response information corresponding to the target consultation intent; a second analysis module, configured to analyze each historical consultation intent and the target consultation intent using a bidirectional long short-term memory network when the first response information includes options, to determine whether there is an option selection intent in the target consultation intent; a first feedback module, configured to feed back the first response information when there is no option selection intent in the target consultation intent; and a second feedback module, configured to determine the second response information based on the target option corresponding to the option selection intent when there is an option selection intent in the target consultation intent, and feed back the second response information.
[0014] According to another aspect of the embodiments of this application, a computer program product is also provided, the computer program product comprising: a computer program, wherein the computer program, when executed by a processor, implements the above-described intelligent response method.
[0015] According to another aspect of the embodiments of this application, an electronic device is also provided, the electronic device including: a memory and a processor, wherein the memory stores a computer program, and the processor is configured to execute the above-described smart response method through the computer program.
[0016] In this embodiment, the target consultation question for the current round is obtained, and a list of historical intents is also obtained. This list includes historical consultation intents corresponding to consultation questions from previous rounds in a multi-round dialogue. Natural language processing algorithms are used to analyze the target consultation question and each historical consultation intent to obtain the target consultation intent corresponding to the target consultation question, and a first response information corresponding to the target consultation intent is determined. If the first response information includes options, a bidirectional long short-term memory network is used to analyze each historical consultation intent and the target consultation intent to determine whether there is an option selection intent in the target consultation intent. If there is no option selection intent in the target consultation intent, the first response information is fed back. If there is an option selection intent in the target consultation intent, a second response information is determined based on the target option corresponding to the option selection intent, and the second response information is fed back. By using natural language processing algorithms to determine multiple consultation intents corresponding to the consultation question, and simultaneously using a bidirectional long short-term memory network to filter multiple consultation intents to determine the final response information, the accuracy of responses can be improved, and the repetition rate of responses can be reduced in different scenarios. This effectively solves the technical problem in related intelligent response schemes where responses are independent in different scenarios, and repetitive responses easily occur in multi-round dialogues, affecting user experience. Attached Figure Description
[0017] The accompanying drawings, which are included to provide a further understanding of this application and form part of this application, illustrate exemplary embodiments and are used to explain this application, but do not constitute an undue limitation of this application. In the drawings:
[0018] Figure 1 This is a schematic diagram of the structure of an optional computer terminal according to an embodiment of this application;
[0019] Figure 2 This is a flowchart illustrating an optional smart response method according to an embodiment of this application;
[0020] Figure 3 This is a schematic diagram of an optional conventional consultation process according to an embodiment of this application;
[0021] Figure 4 This is a schematic diagram of an optional smart response method consultation process according to an embodiment of this application;
[0022] Figure 5 This is a schematic diagram of an optional smart response device according to an embodiment of this application. Detailed Implementation
[0023] To enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present application, and not all embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative effort should fall within the scope of protection of the present application.
[0024] It should be noted that the terms "first," "second," etc., used in the specification, claims, and drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this application described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.
[0025] To better understand the embodiments of this application, the following is a translation and explanation of some nouns or terms that appear in the description of the embodiments of this application:
[0026] Natural Language Processing (NLP) is a set of theories and methods that enable effective communication between humans and computers using natural language.
[0027] Long Short-Term Memory (LSTM) network: It is a type of time-recurrent neural network that can solve the long-term dependency problem of general recurrent neural networks. It is composed of a forward LSTM and a backward LSTM.
[0028] Example 1
[0029] According to an embodiment of this application, an intelligent response method is provided. It should be noted that the steps shown in the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions. Furthermore, although a logical order is shown in the flowchart, in some cases, the steps shown or described may be executed in a different order than that shown here.
[0030] The methods and embodiments provided in this application can be executed on mobile terminals, computer terminals, or similar computing devices. Figure 1A hardware block diagram of a computer terminal (or mobile device) for implementing an intelligent response method is shown. Figure 1 As shown, the computer terminal 10 (or mobile device 10) may include one or more processors 102 (shown as 102a, 102b, ..., 102n in the figure) 102 (processor 102 may include, but is not limited to, a microprocessor MCU or a programmable logic device FPGA, etc.), a memory 104 for storing data, and a transmission device 106 for communication functions. In addition, it may also include: a display, an input / output interface (I / O interface), a universal serial bus (USB) port (which may be included as one of the ports of a BUS bus), a network interface, a power supply, and / or a camera. Those skilled in the art will understand that... Figure 1 The structure shown is for illustrative purposes only and does not limit the structure of the aforementioned electronic device. For example, computer terminal 10 may also include... Figure 1 The more or fewer components shown, or having the same Figure 1 The different configurations shown.
[0031] It should be noted that the aforementioned one or more processors 102 and / or other data processing circuits are generally referred to herein as "data processing circuits". These data processing circuits may be embodied, in whole or in part, in software, hardware, firmware, or any other combination thereof. Furthermore, the data processing circuits may be a single, independent processing module, or may be integrated, in whole or in part, into any other element within the computer terminal 10 (or mobile device). As involved in the embodiments of this application, the data processing circuits serve as a processor control mechanism (e.g., selection of a variable resistor termination path connected to an interface).
[0032] The memory 104 can be used to store software programs and modules of application software, such as the program instructions / data storage device corresponding to the intelligent response method in this embodiment. The processor 102 executes various functional applications and data processing by running the software programs and modules stored in the memory 104, thereby implementing the above-mentioned vulnerability detection method for the application. The memory 104 may include high-speed random access memory, and may also include non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some instances, the memory 104 may further include memory remotely located relative to the processor 102, and these remote memories can be connected to the computer terminal 10 via a network. Examples of such networks include, but are not limited to, the Internet, corporate intranets, local area networks, mobile communication networks, and combinations thereof.
[0033] The transmission device 106 is used to receive or send data via a network. Specific examples of the network described above may include a wireless network provided by the communication provider of the computer terminal 10. In one example, the transmission device 106 includes a Network Interface Controller (NIC), which can connect to other network devices via a base station to communicate with the Internet. In another example, the transmission device 106 may be a Radio Frequency (RF) module, used for wireless communication with the Internet.
[0034] The display can be, for example, a touchscreen liquid crystal display (LCD) that allows the user to interact with the user interface of the computer terminal 10 (or mobile device).
[0035] Under the above operating environment, this application provides an intelligent response method, such as... Figure 2 As shown, the method includes the following steps:
[0036] Step S202: Obtain the target consultation question for the current round and obtain the historical intent list, wherein the historical intent list includes: the historical consultation intents corresponding to the historical consultation questions in each round before the current round in the multi-round dialogue;
[0037] Step S204: Analyze the target consultation question and each historical consultation intent using a natural language processing algorithm to obtain the target consultation intent corresponding to the target consultation question, and determine the first response information corresponding to the target consultation intent;
[0038] Step S206: If the first response information includes options, use a bidirectional long short-term memory network to analyze each historical consultation intent and the target consultation intent to determine whether there is an option selection intent in the target consultation intent.
[0039] Step S208: If there is no option selection intent in the target consultation intent, the first response information is fed back;
[0040] Step S210: If the target consultation intent contains an option selection intent, determine the second response information based on the target option corresponding to the option selection intent, and provide feedback on the second response information.
[0041] The following section explains each step of the intelligent response method in conjunction with a specific implementation process.
[0042] As an optional implementation, the target consultation question and various historical consultation intentions can be analyzed using natural language processing algorithms in the following ways to obtain the target consultation intention corresponding to the target consultation question: the target consultation question is analyzed using semantic analysis algorithms to obtain the target semantic information corresponding to the target consultation question; and various historical consultation intentions and target semantic information are analyzed using a bidirectional long short-term memory network to obtain the target consultation intention corresponding to the target consultation question.
[0043] In the bidirectional long short-term memory network, the forward LSTM and backward LSTM operate independently, but their outputs are combined to form a single representation of the input sequence. The forward LSTM processes the input sequence from left to right, while the backward LSTM processes the sequence from right to left. The outputs of these two layers can be connected and fed back to a fully connected layer for classification or other downstream tasks. The bidirectional long short-term memory network can process the past and future context of the input sequence simultaneously, thereby better capturing the semantic and grammatical features of the language. In addition, the bidirectional long short-term memory network can be trained using techniques such as attention mechanisms, enabling it to focus on specific parts of the input sequence during training and inference.
[0044] For example, the forward LSTM is fed with "I", "love", and "you" in sequence, resulting in three vectors {hL0, hL1, hL2}. The backward LSTM is fed with "you", "love", and "I" in sequence, resulting in three vectors {hR0, hR1, hR2}. The forward and backward hidden vectors are then concatenated to obtain {[hL0, hR2], [hL1, hR1], [hL2, hR0]}, or {h0, h1, h2}. For sentiment classification tasks, the sentence usually used is [hL2, hR2], because it contains all the information from both the forward and backward processes.
[0045] Then, the first response information corresponding to the target consultation intent can be determined in the following way: determine the first consultation scenario corresponding to the target consultation intent; determine the first response information corresponding to the first consultation scenario from the preset mapping relationship table, wherein the mapping relationship table stores the mapping relationship between multiple consultation scenarios and multiple response information.
[0046] It is understandable that the first consultation scenario can be multiple consultation scenarios such as discount packages and inquiries about costs. Assuming that the determined first consultation scenario is discount packages, the corresponding first response information can be found from the preset mapping relationship table: 1. mobile phone, 2. broadband.
[0047] After determining the first response information corresponding to the target consultation intent, the first response information can be fed back even if it does not include options.
[0048] If, after a user selects a mobile phone option, the first response is "5G Enjoy Package" but does not include the user's desired option, the user can automatically enter the desired option, such as "Package Tier". In this case, the first response will continue to be "5G Enjoy Package".
[0049] Optionally, the second response information can be determined based on the target option corresponding to the option selection intent in the following manner: determine the target option corresponding to the option selection intent from the options in the first response information, and determine the second consultation scenario corresponding to the target option; determine the second response information corresponding to the second consultation scenario from the mapping relationship table.
[0050] Optionally, the second response information can be fed back in the following way: the second response information and the other options in the first response information, except for the target option, are fed back simultaneously.
[0051] After providing feedback and response information, the target consultation intent can be added to the historical intent list.
[0052] It should be noted that adding the target consultation intent hit in this consultation to the historical intent list after the feedback response information is provided can further improve the intelligence of the intelligent question-and-answer process for the next consultation understanding.
[0053] Figure 3 A schematic diagram of an alternative traditional solution consultation process is shown, such as... Figure 3 As shown, in the consultation process of traditional solutions, after understanding that the user's consultation is biased towards mobile phones, when entering the package tier scenario, the user still needs to make a choice again when it comes to the options of mobile phones and broadband, which greatly affects the user experience. Figure 4 A schematic diagram of an optional intelligent response method consultation process is shown, such as... Figure 4 As shown, the optimized intelligent response method in this application embodiment can optimize the semantic understanding system. When it is understood that the user's inquiry is biased towards mobile phones, the same selection does not need to be made again when entering the package level scenario, which greatly optimizes the intelligence of the intelligent question and answer process, saves the complexity of human operation, and improves the user experience.
[0054] In this embodiment, the target consultation question for the current round is obtained, and a list of historical intents is also obtained. This list includes historical consultation intents corresponding to consultation questions from previous rounds in a multi-round dialogue. Natural language processing algorithms are used to analyze the target consultation question and each historical consultation intent to obtain the target consultation intent corresponding to the target consultation question, and a first response information corresponding to the target consultation intent is determined. If the first response information includes options, a bidirectional long short-term memory network is used to analyze each historical consultation intent and the target consultation intent to determine whether there is an option selection intent in the target consultation intent. If there is no option selection intent in the target consultation intent, the first response information is fed back. If there is an option selection intent in the target consultation intent, a second response information is determined based on the target option corresponding to the option selection intent, and the second response information is fed back. By using natural language processing algorithms to determine multiple consultation intents corresponding to the consultation question, and simultaneously using a bidirectional long short-term memory network to filter multiple consultation intents to determine the final response information, the accuracy of responses can be improved, and the repetition rate of responses can be reduced in different scenarios. This effectively solves the technical problem in related intelligent response schemes where responses are independent in different scenarios, and repetitive responses easily occur in multi-round dialogues, affecting user experience.
[0055] Example 2
[0056] According to an embodiment of this application, an intelligent response device for implementing the intelligent response method in Embodiment 1 is also provided, such as... Figure 5 As shown, the intelligent response device includes at least: an acquisition module 51, a first analysis module 52, a second analysis module 53, a first feedback module 54, and a second feedback module 55, wherein:
[0057] The acquisition module 51 can acquire the target consultation question for the current round and acquire the historical intent list, which includes the historical consultation intents corresponding to the historical consultation questions in each round before the current round in the multi-round dialogue.
[0058] The first analysis module 52 can use natural language processing algorithms to analyze the target consultation question and various historical consultation intentions, obtain the target consultation intention corresponding to the target consultation question, and determine the first response information corresponding to the target consultation intention;
[0059] The second analysis module 53 can analyze each historical consultation intent and the target consultation intent using a bidirectional long short-term memory network when the first response information includes options, to determine whether there is an option selection intent in the target consultation intent;
[0060] The first feedback module 54 can provide the first response information when there is no option to select intent in the target consultation intent;
[0061] When there is an option selection intent in the target consultation intent, the second feedback module 55 can determine the second response information based on the target option corresponding to the option selection intent and then feed back the second response information.
[0062] The functions of each module of the intelligent response device are explained below in conjunction with the specific implementation process.
[0063] As an optional implementation, the first analysis module can use natural language processing algorithms to analyze the target consultation question and various historical consultation intentions to obtain the target consultation intention corresponding to the target consultation question in the following ways: analyze the target consultation question using semantic analysis algorithms to obtain the target semantic information corresponding to the target consultation question; and analyze various historical consultation intentions and target semantic information using a bidirectional long short-term memory network to obtain the target consultation intention corresponding to the target consultation question.
[0064] In the bidirectional long short-term memory network, the forward LSTM and backward LSTM operate independently, but their outputs are combined to form a single representation of the input sequence. The forward LSTM processes the input sequence from left to right, while the backward LSTM processes the sequence from right to left. The outputs of these two layers can be connected and fed back to a fully connected layer for classification or other downstream tasks. The bidirectional long short-term memory network can process the past and future context of the input sequence simultaneously, thereby better capturing the semantic and grammatical features of the language. In addition, the bidirectional long short-term memory network can be trained using techniques such as attention mechanisms, enabling it to focus on specific parts of the input sequence during training and inference.
[0065] For example, the forward LSTM is fed with "I", "love", and "you" in sequence, resulting in three vectors {hL0, hL1, hL2}. The backward LSTM is fed with "you", "love", and "I" in sequence, resulting in three vectors {hR0, hR1, hR2}. The forward and backward hidden vectors are then concatenated to obtain {[hL0, hR2], [hL1, hR1], [hL2, hR0]}, or {h0, h1, h2}. For sentiment classification tasks, the sentence usually used is [hL2, hR2], because it contains all the information from both the forward and backward processes.
[0066] Subsequently, the first analysis module can determine the first response information corresponding to the target consultation intent in the following way: determine the first consultation scenario corresponding to the target consultation intent; determine the first response information corresponding to the first consultation scenario from a preset mapping relationship table, wherein the mapping relationship table stores the mapping relationship between multiple consultation scenarios and multiple response information.
[0067] It is understandable that the first consultation scenario can be multiple consultation scenarios such as discount packages and inquiries about costs. Assuming that the determined first consultation scenario is discount packages, the corresponding first response information can be found from the preset mapping relationship table: 1. mobile phone, 2. broadband.
[0068] After determining the first response information corresponding to the target consultation intent, the first feedback module can provide the first response information even if the first response information does not include options.
[0069] If, after a user selects a mobile phone option, the first response is "5G Enjoy Package" but does not include the user's desired option, the user can automatically enter the desired option, such as "Package Tier". In this case, the first response will continue to be "5G Enjoy Package".
[0070] Optionally, the second feedback module can determine the second response information based on the target option corresponding to the option selection intent in the following way: determine the target option corresponding to the option selection intent from the options in the first response information, and determine the second consultation scenario corresponding to the target option; determine the second response information corresponding to the second consultation scenario from the mapping relationship table.
[0071] Optionally, the second feedback module can provide the second response information in the following way: simultaneously provide the second response information and the other options in the first response information, excluding the target option.
[0072] After providing feedback and response information, the target consultation intent can be added to the historical intent list.
[0073] It should be noted that adding the target consultation intent hit in this consultation to the historical intent list after the feedback response information is provided can further improve the intelligence of the intelligent question-and-answer process for the next consultation understanding.
[0074] Figure 3 A schematic diagram of an alternative traditional solution consultation process is shown, such as... Figure 3 As shown, in the consultation process of traditional solutions, after understanding that the user's consultation is biased towards mobile phones, when entering the package tier scenario, the user still needs to make a choice again when it comes to the options of mobile phones and broadband, which greatly affects the user experience. Figure 4 A schematic diagram of an optional intelligent response method consultation process is shown, such as... Figure 4 As shown, the optimized intelligent response method in this application embodiment can optimize the semantic understanding system. When it is understood that the user's inquiry is biased towards mobile phones, the same selection does not need to be made again when entering the package level scenario, which greatly optimizes the intelligence of the intelligent question and answer process, saves the complexity of human operation, and improves the user experience.
[0075] It should be noted that each module in the intelligent response device in this application embodiment corresponds one-to-one with each implementation step of the intelligent response method in embodiment 1. Since embodiment 1 has been described in detail, some details not shown in this embodiment can be referred to embodiment 1, and will not be elaborated further here.
[0076] Example 3
[0077] According to an embodiment of this application, a computer program product is also provided, which includes a computer program, wherein the computer program implements the above-described intelligent response method when executed by a processor.
[0078] According to an embodiment of this application, a non-volatile storage medium is also provided, which includes a stored computer program, wherein the device containing the non-volatile storage medium executes the smart response method in Embodiment 1 by running the computer program.
[0079] According to an embodiment of this application, a processor is also provided for running a computer program, wherein the computer program executes the smart response method in embodiment 1 during runtime.
[0080] According to an embodiment of this application, an electronic device is also provided, comprising: a memory and a processor, wherein the memory stores a computer program, and the processor is configured to execute the smart response method of embodiment 1 through the computer program.
[0081] Specifically, the computer program executes the following steps during runtime: It obtains the target consultation question for the current round and a list of historical intents, whereby the list includes historical consultation intents corresponding to historical consultation questions from previous rounds in a multi-round dialogue; it analyzes the target consultation question and each historical consultation intent using a natural language processing algorithm to obtain the target consultation intent corresponding to the target consultation question and determines the first response information corresponding to the target consultation intent; if the first response information includes options, it analyzes each historical consultation intent and the target consultation intent using a bidirectional long short-term memory network to determine whether there is an option selection intent in the target consultation intent; if there is no option selection intent in the target consultation intent, it feeds back the first response information; if there is an option selection intent in the target consultation intent, it determines the second response information based on the target option corresponding to the option selection intent and feeds back the second response information.
[0082] The sequence numbers of the above embodiments are for descriptive purposes only and do not represent the superiority or inferiority of the embodiments.
[0083] In the above embodiments of this application, the descriptions of each embodiment have different focuses. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions of other embodiments.
[0084] In the several embodiments provided in this application, it should be understood that the disclosed technical content can be implemented in other ways. The device embodiments described above are merely illustrative; for example, the division of units can be a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the displayed or discussed mutual couplings, direct couplings, or communication connections may be through some interfaces; indirect couplings or communication connections between units or modules may be electrical or other forms.
[0085] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.
[0086] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.
[0087] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as a USB flash drive, read-only memory (ROM), random access memory (RAM), portable hard drive, magnetic disk, or optical disk.
[0088] The above are merely preferred embodiments of this application. It should be noted that those skilled in the art can make various improvements and modifications without departing from the principles of this application, and these improvements and modifications should also be considered within the scope of protection of this application.
Claims
1. A smart response method, characterized in that, include: Obtain the target consultation question for the current round and obtain a list of historical intents, wherein the list of historical intents includes: historical consultation intents corresponding to historical consultation questions in each round before the current round in multi-round dialogue; The target consultation question and each of the historical consultation intentions are analyzed using natural language processing algorithms to obtain the target consultation intention corresponding to the target consultation question, and to determine the first response information corresponding to the target consultation intention. If the first response information includes options, a bidirectional long short-term memory network is used to analyze each of the historical consultation intentions and the target consultation intention to determine whether there is an option selection intention in the target consultation intention; If the target consultation intent does not include an option selection intent, the first response information is returned. If the target consultation intent contains an option selection intent, the second response information is determined based on the target option corresponding to the option selection intent, and the second response information is fed back.
2. The method according to claim 1, characterized in that, The target consultation question and each of the historical consultation intentions are analyzed using natural language processing algorithms to obtain the target consultation intention corresponding to the target consultation question, including: The target consultation question is analyzed using a semantic analysis algorithm to obtain the target semantic information corresponding to the target consultation question; By using a bidirectional long short-term memory network to analyze each of the historical consultation intentions and the target semantic information, the target consultation intention corresponding to the target consultation question can be obtained.
3. The method according to claim 1, characterized in that, Determining the first response information corresponding to the stated consultation intent includes: Determine the first consultation scenario corresponding to the stated target consultation intent; The first response information corresponding to the first consultation scenario is determined from a preset mapping relationship table, wherein the mapping relationship table stores the mapping relationship between multiple consultation scenarios and multiple response information.
4. The method according to claim 1, characterized in that, After determining the first response information corresponding to the target consultation intent, the method further includes: If the first response information does not include options, then the first response information is returned.
5. The method according to claim 3, characterized in that, Based on the selected options, the target option corresponding to the intent is chosen to determine the second response information, including: From the options in the first response information, determine the target option corresponding to the option selection intent, and determine the second consultation scenario corresponding to the target option; The second response information corresponding to the second consultation scenario is determined from the mapping table.
6. The method according to claim 5, characterized in that, The second response information includes: The second response information and all other options in the first response information except for the target option are fed back simultaneously.
7. The method according to claim 1, characterized in that, After providing the response information, the method further includes: Add the target consultation intent to the historical intent list.
8. A smart response device, characterized in that, include: The acquisition module is used to acquire the target consultation question for the current round and acquire a list of historical intents, wherein the list of historical intents includes: the historical consultation intents corresponding to the historical consultation questions in each round before the current round in the multi-round dialogue; The first analysis module is used to analyze the target consultation question and each of the historical consultation intentions using natural language processing algorithms, to obtain the target consultation intention corresponding to the target consultation question, and to determine the first response information corresponding to the target consultation intention; The second analysis module is used to analyze each of the historical consultation intentions and the target consultation intentions using a bidirectional long short-term memory network when the first response information includes options, to determine whether there is an option selection intention in the target consultation intention; The first feedback module is used to provide feedback on the first response information when there is no option selection intent in the target consultation intent; The second feedback module is used to determine the second response information based on the target option corresponding to the option selection intention when there is an option selection intention in the target consultation intention, and to feed back the second response information.
9. A computer program product, characterized in that, include: A computer program, wherein when executed by a processor, the computer program implements the intelligent response method according to any one of claims 1 to 7.
10. An electronic device, characterized in that, include: A memory and a processor, wherein the memory stores a computer program, and the processor is configured to execute the smart response method according to any one of claims 1 to 7 through the computer program.