Voice number inquiry method and device

Abstract

The present invention discloses a method and device for voice directory inquiry, wherein a voice directory inquiry method comprises: in response to a voice robot receiving voice directory inquiry information input by a user, extracting a first organization name from the voice directory inquiry information, if the first organization name does not exist in a preset set of multiple organization names, judging whether there is a second organization name similar to the first organization name in the set of multiple organization names; if there is a second organization name similar to the first organization name in the set of multiple organization names, confirming to the user whether the first organization name and the second organization name are semantically identical; if a semantically identical instruction from the user is received, feeding back the number information corresponding to the second organization name to the user, enriching the organization name set according to the semantically identical instruction from the user, thereby realizing self-learning. By using voice interaction, the difficulties of inaccurate directory inquiry information and digitalization and intelligence of directory inquiry services are solved.

Description

Technical Field

[0001] The invention belongs to the technical field of intelligent voice number query, and in particular relates to a voice number query method and device. Background Art

[0002] There are two common ways to look up numbers: one is to search online directly. Users search for numbers through online search engines. The returned results are very broad, making it difficult to quickly locate valuable information. The search efficiency is low, and much of the information is not authentic. The other is to call a related hotline (such as 114) and ask the agent for the name of the organization they want to look up.

[0003] If the agent is human, they need to manually search the organization directory and then tell the user the phone records. This manual agent processing method is inefficient, and only one agent can connect with one user at a time. If 24-hour service is required, they have to work in shifts, so the cost is relatively high. If the agent is a robot, it can quickly broadcast the query results to the user after recognizing the voice, but it cannot flexibly handle various abbreviations of the same organization and does not have the ability to self-learn, which will greatly reduce the experience of voice interaction with users. Summary of the invention

[0004] The embodiments of the present invention provide a voice directory query method and device, which are used to solve at least one of the above technical problems.

[0005] In a first aspect, an embodiment of the present invention provides a voice directory inquiry method, comprising: in response to a voice robot receiving voice directory inquiry information input by a user, extracting a first institution name from the voice directory inquiry information, and querying whether the first institution name exists in a preset set of multiple institution names; if the first institution name exists in the set of multiple institution names, feeding back number information corresponding to the first institution name to the user; if the first institution name does not exist in any of the preset sets of multiple institution names, the voice robot determines whether there is a second institution name similar to the first institution name in the set of multiple institution names; if there is a second institution name similar to the first institution name in the set of multiple institution names, confirming with the user whether the first institution name and the second institution name are semantically identical; if a semantic identical instruction from the user is received, the voice robot feeds back the number information corresponding to the second institution name to the user.

[0006] In a second aspect, an embodiment of the present invention provides a voice directory inquiry device, comprising: an extraction program module, configured to extract a first institution name from the voice directory inquiry information in response to a voice robot receiving voice directory inquiry information input by a user, and query whether the first institution name exists in a preset set of multiple institution names; an existence processing program module, configured to feedback number information corresponding to the first institution name to the user if the first institution name exists in the set of multiple institution names; a similarity judgment program module, configured to, if the first institution name does not exist in the preset set of multiple institution names, the voice robot judges whether there is a second institution name similar to the first institution name in the set of multiple institution names; a confirmation program module, configured to, if there is a second institution name similar to the first institution name in the set of multiple institution names, confirm to the user whether the first institution name and the second institution name are semantically identical; a feedback program module, configured to, if a semantically identical instruction from the user is received, the voice robot feedbacks the number information corresponding to the second institution name to the user.

[0007] In a third aspect, an electronic device is provided, comprising: at least one processor, and a memory communicatively connected to the at least one processor, wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the steps of the method described in the first aspect.

[0008] The method and device of the present application provide a solution that uses voice recognition / natural language understanding / semantic similarity matching and other technologies to solve the problems of inaccurate directory information, the inability of voice robots to flexibly handle abbreviations, and the difficulty of digitalization and intelligence of directory services. Using this voice directory query method, users can query the number of a certain institution more quickly and accurately. BRIEF DESCRIPTION OF THE DRAWINGS

[0009] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0010] Figure 1 A flowchart of a voice number query method provided by one embodiment of the present invention;

[0011] Figure 2 A flowchart of another voice number query method provided by one embodiment of the present invention;

[0012] Figure 3A flowchart of another voice number query method provided by an embodiment of the present invention;

[0013] Figure 4 A flowchart of another voice number query method provided by an embodiment of the present invention;

[0014] Figure 5 A flowchart of another voice number query method provided by an embodiment of the present invention;

[0015] Figure 6 A flowchart of another voice number query method provided by an embodiment of the present invention;

[0016] Figure 7 A flowchart of a storage structure of a directory query scheme provided by an embodiment of the present invention;

[0017] Figure 8 A solution architecture diagram provided for an embodiment of the present invention;

[0018] Fig. 9 A flowchart of an actual conversation example of a voice directory inquiry solution provided by an embodiment of the present invention;

[0019] Fig.10 A software implementation flow chart provided for an embodiment of the present invention;

[0020] Fig.11 A block diagram of a voice directory inquiry device provided by one embodiment of the present invention;

[0021] Fig.12 It is a schematic diagram of the structure of an electronic device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0022] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0023] Please refer to Figure 1 , which shows a flow chart of an embodiment of the voice directory inquiry method of the present application. The voice directory inquiry method of this embodiment can be applied to terminals with communication or real-time voice conversation functions, such as smart phones, tablets, computers, etc. For example, when it is necessary to find the number of a certain organization, the above-mentioned devices can be used to directly search through a voice robot without time or geographical restrictions, which is convenient and fast.

[0024] like Figure 1As shown, in step 101, in response to the voice robot receiving the voice directory information input by the user, a first organization name is extracted from the voice directory information, and a query is made as to whether the first organization name exists in a preset set of multiple organization names;

[0025] In step 102, if the first institution name exists in the set of multiple institution names, the number information corresponding to the first institution name is fed back to the user;

[0026] In step 103, if the first institution name does not exist in the preset multiple institution name sets, the voice robot determines whether there is a second institution name similar to the first institution name in the multiple institution name sets;

[0027] In step 104, if there is a second institution name similar to the first institution name in the set of multiple institution names, confirming with the user whether the first institution name and the second institution name are semantically identical;

[0028] In step 105, if the voice robot receives a semantically identical instruction from the user, it will feed back the number information corresponding to the second institution name to the user.

[0029] In this embodiment, for step 101, the voice directory inquiry device extracts the first institution name from the voice directory inquiry information in response to the voice robot receiving the voice directory inquiry information input by the user, and queries whether the first institution name exists in a preset set of multiple institution names. For example, the voice information input by the user is an instruction of "query the telephone number of Shanghai Jiao Tong University", and the voice robot extracts the first institution name "Shanghai Jiao Tong University" and searches from the preset database institution name set whether there is information on institution names related to "Shanghai Jiao Tong University".

[0030] Afterwards, for step 102, if the first institution name exists in the set of multiple institution names, the number information corresponding to the first institution name will be fed back to the user. For example, the voice robot finds that "Shanghai Jiao Tong University" exists in the institution set related to "Shanghai Jiao Tong University", and the voice robot broadcasts the number information of "Shanghai Jiao Tong University" to the user.

[0031] Then, for step 103, if the first institution name does not exist in the preset multiple institution name sets, the voice robot determines whether there is a second institution name similar to the first institution name in the multiple institution name sets. For example, if the user enters "Shanghai Jiaotong University", the voice robot does not find the institution "Shanghai Jiaotong University" in the "Shanghai Jiaotong University" institution set. The voice robot searches for an institution name similar to "Shanghai Jiaotong University" based on the judgment and analysis.

[0032] Afterwards, for step 104, if there is a second institution name similar to the first institution name in the set of multiple institution names, confirm with the user whether the first institution name and the second institution name are semantically identical. For example, the voice machine finds a set of "Shanghai Jiao Tong University" institution names similar to "Shanghai Jiao Tong University", and the voice robot announces to the user whether the institution being queried is "Shanghai Jiao Tong University".

[0033] Finally, for step 105, if the user's semantically identical instruction is received, the voice robot will feed back the number information corresponding to the second institution name to the user. For example, if the voice robot receives the user's "yes" instruction, the voice robot will broadcast the number information of the "Shanghai Jiaotong University" institution found to the user. In some optional embodiments, the voice robot can enrich the institution name collection according to the user's confirmation feedback, thereby achieving self-learning. For example, after the user feedbacks semantically identical or semantically similar, the semantically identical or similar institution name (such as "Shanghai Jiaotong University") queried by the user can be added to the institution name collection to enrich the institution name collection and achieve self-learning. In some other embodiments, if the user inquires about "Jiaotong University" and finds multiple institution names with the same or similar meanings as "Jiaotong University", such as "Beijing Jiaotong University", "Shanghai Jiaotong University", etc., at this time, even if the user later selects or confirms "Shanghai Jiaotong University", it is not appropriate to directly add "Jiaotong University" to the institution name set of "Shanghai Jiaotong University", because the abbreviation may also correspond to "Beijing Jiaotong University" in subsequent institution name queries. If it is directly added to the institution name set of "Shanghai Jiaotong University", it may cause the abbreviation "Jiaotong University" to be directly corresponded to "Shanghai Jiaotong University" when querying in the future, which will seriously affect the accuracy of the query and affect the user experience.

[0034] The method of this embodiment uses the form of voice interaction between the user and the voice robot to search for similar or identical organization names in the set according to a preset organization name set, and then feeds back the number information corresponding to the organization name to the user. Furthermore, if the abbreviation of the organization name is not in the set, it can be stored in the corresponding full name set of the organization name, which solves the time limit of manual service being difficult to ensure 24-hour on-the-job service, while avoiding labor costs and improving work efficiency.

[0035] Further references Figure 2 , which shows a flowchart of another voice number query method provided by an embodiment of the present invention, the flowchart is mainly Figure 1 A flowchart of the steps further defined in step 103 of the process "the voice robot determines whether there is a second institution name similar to the first institution name in the set of multiple institution names".

[0036] like Figure 2 As shown, in step 201, the voice robot calculates the semantic similarity between the abbreviations in the set of multiple institution names and the first institution name;

[0037] In step 202, if there is a second institution name in at least one institution name set whose similarity to the first institution name is greater than or equal to the semantic similarity threshold, it is confirmed that there is a second institution name similar to the first institution name in the plurality of institution name sets.

[0038] In this embodiment, for step 201, the voice robot calculates the semantic similarity between the abbreviations in the set of multiple organization names and the first organization name. For example, the voice robot calculates the semantic similarity between the abbreviations in the originally set organization name set and similar organization names or abbreviations that the user needs to query.

[0039] Finally, for step 202, the voice robot searches for the result. If there is a second institution name in at least one institution name set whose similarity with the first institution name is greater than or equal to the semantic similarity threshold, then it is confirmed that there is a second institution name similar to the first institution name in the multiple institution name sets. For example, the voice robot finds the institution name "Shanghai Jiao Tong University" in the institution name set pre-stored in the database, and calculates through a pre-set threshold that the similarity between the institution name "Shanghai Jiao Tong University" that the user needs to find and the existing institution names in the "Shanghai Jiao Tong University" institution name set is greater than or equal to the preset similarity threshold, then it is confirmed that there is an institution name in the "Shanghai Jiao Tong University" institution name set that is similar to the institution name the user wants to query.

[0040] The method of this embodiment uses a preset semantic similarity threshold to determine whether there are any organization names similar to the organization name the user needs to find among the organization names already stored in the system. Even if there are no identical organization names, organization names similar to the organization name the user needs to find can be quickly found.

[0041] In some optional embodiments, there are multiple second institution names whose similarity with the first institution name is greater than or equal to the semantic similarity threshold, and only the second institution name with the highest similarity is obtained. For example, the results found by the voice robot are "Shanghai Jiaotong University", "Shanghai Jiaotong University", "Jiaotong University", and "Jiaotong University". The preset semantic similarity threshold is calculated with the "Shanghai Jiaotong University" that the user needs to find, and the institution name with the highest similarity is obtained. For example, "Shanghai Jiaotong University" has the highest similarity with the "Shanghai Jiaotong University" that the user needs to find, so only "Shanghai Jiaotong University" is obtained, thereby avoiding the need for users to confirm too many similar institution names, which greatly affects the user's interactive experience. The method of this embodiment uses a voice robot to find multiple institution names that are similar to the institution name that the user needs to find, and according to the preset threshold, only the institution name with the highest similarity is taken, which solves the problem of multiple similar institution names being difficult to select.

[0042] In some optional embodiments, each of the multiple institution name sets has a complete institution name that uniquely corresponds to the institution name set, and the institution name set includes the abbreviation of the complete institution name and the full name of the complete institution name. For example, the institution name of "Shanghai Jiaotong University" is changed to include "Shanghai Jiaotong University", "Shanghai Jiaotong University", "Shanghai Jiaotong University", etc. in the set, and each set must have a complete institution name such as "Shanghai Jiaotong University", where the institution name set of "Shanghai Jiaotong University" can also include its abbreviations "Shanghai Jiaotong University", "Shanghai Jiaotong University", etc. The method of this embodiment requires that the complete institution name must be in the pre-set institution name set, so that the voice robot can find the institution name set that the user needs to find faster and more accurately.

[0043] Please refer to Figure 3 , which shows a flowchart of another voice number query method provided by an embodiment of the present invention, the flowchart is mainly Figure 1 A flowchart of the steps further defined in step 105 of the process "the voice robot feeds back the number information corresponding to the second institution name to the user".

[0044] like Figure 3 As shown, in step 301, a complete institution name of an institution name set corresponding to the second institution name is obtained;

[0045] In step 302, the number information corresponding to the second organization name and the corresponding complete organization name are fed back to the user.

[0046] In this embodiment, for step 301, the voice robot obtains the complete institution name of the institution name set corresponding to the second institution name. For example, the robot obtains the corresponding institution name set such as {"Shanghai Jiao Tong University", "Shanghai Jiao Tong University", "Shanghai Jiao Tong University"}, etc., and the voice robot obtains the complete institution name such as "Shanghai Jiao Tong University" in the set.

[0047] Then, for step 302, the voice robot will feed back the number information corresponding to the second institution name and the corresponding complete institution name to the user. For example, the voice robot will broadcast the complete institution name "Shanghai Jiao Tong University" and the number information "021-****" corresponding to "Shanghai Jiao Tong University" to the user.

[0048] The method of this embodiment uses a voice robot to find the set of organization names that the user needs to find, and confirm with the user whether it is the organization. If it is the organization, the full name of the organization in the set and the corresponding number information are broadcast to the user, so that the user can understand the full name of the organization and the corresponding telephone number more clearly. By broadcasting the entire organization name, the user is given an opportunity to confirm again whether the broadcasted organization name is the organization name the user wants to query.

[0049] Please refer to Figure 4 , which shows a flowchart of another voice number query method provided by an embodiment of the present invention, the flowchart is mainly Figure 1 A flowchart of the steps further defined in step 104 of the process "If there is a second institution name similar to the first institution name in the set of multiple institution names, confirm with the user whether the first institution name and the second institution name are semantically identical."

[0050] like Figure 4 As shown, in step 401, if there is and only one second institution name in the set of multiple institution names that is similar to the first institution name, confirm with the user whether the first institution name and the second institution name are semantically identical;

[0051] In step 402, if there are multiple second institution names in the set of multiple institution names that are similar to the first institution name, confirm with the user whether any of the multiple second institution names are semantically identical.

[0052] In this embodiment, for step 401, if there is only one second institution name in the set of multiple institution names that is similar to the first institution name, confirm with the user whether the first institution name and the second institution name are semantically identical. For example, in a set of multiple institution names such as "Shanghai Jiao Tong University", "Shanghai Jiao Tong", "Shanghai Jiao Tong University", etc., among which "Shanghai Jiao Tong University" is similar to the "Shanghai Jiao Tong University" that the user needs to find, the voice robot confirms with the user whether it is "Shanghai Jiao Tong University".

[0053] Then, for step 402, if there are multiple second institution names in the set of multiple institution names that are similar to the first institution name, confirm with the user whether any of the multiple second institution names are semantically identical. For example, the user inputs the institution name "Jiaotong University" to search for, and the voice robot finds multiple institution names "Shanghai Jiaotong University", "Beijing Jiaotong University", and "Nanjing Jiaotong University", and asks the user whether it is: "Shanghai Jiaotong University", "Beijing Jiaotong University" or "Nanjing Jiaotong University".

[0054] The method of this embodiment uses a voice robot to find a set of one or more organization names that are similar to the organization name that the user needs to find, and confirms with the user whether it is the organization name; this way of confirming with the user ensures that the voice robot will not make mistakes when searching for the organization name that the user needs.

[0055] Please refer to Figure 5 , which shows a flowchart of another voice number query method provided by an embodiment of the present invention, the flowchart is mainly Figure 4 A flowchart of the steps further defined in the process.

[0056] like Figure 5 As shown, in step 501, if the user's semantically identical instruction is received, the voice robot feeds back the number information corresponding to the second institution name to the user;

[0057] In step 502, if there is only one second organization name that is similar to the first organization name, and the user confirms that the semantics are the same, the first organization name is stored in the organization name set corresponding to the second organization name to enrich the organization name set to achieve self-learning of the voice robot, and the number information corresponding to the second organization name is fed back to the user.

[0058] In this embodiment, for step 501, the voice robot receives the semantically identical instruction from the user, and the voice robot feeds back the number information corresponding to the second institution name to the user. For example, the voice robot receives that the institution name searched by the user is "Shanghai Jiao Tong University", and the voice robot directly broadcasts the number information corresponding to "Shanghai Jiao Tong University" to the user.

[0059] Then, for step 502, if there is only one second institution name that is similar to the first institution name, and the user confirms that the semantics are the same, the first institution name is stored in the institution name set corresponding to the second institution name, and the number information corresponding to the second institution name is fed back to the user. For example, the institution name "Shanghai Jiao Tong University" found by the voice robot is most similar to the institution name "Shanghai Jiao Tong University" that the user needs to find, and the user also confirms that it is "Shanghai Jiao Tong University". The voice robot stores the abbreviation "Shanghai Jiao Tong University" in the "Shanghai Jiao Tong University" set, and at the same time broadcasts the number information of "Shanghai Jiao Tong University" to the user.

[0060] The method of this embodiment uses the voice robot and the user to confirm whether it is the organization name that the voice robot is looking for. If the abbreviation of the organization name that the user needs to find does not exist in the database, the voice robot stores the new abbreviation of the organization name in the corresponding organization name set, and broadcasts the number information corresponding to the organization name to the user. This method can more accurately find the organization name number required by the user. Furthermore, in the case where there is only one organization name that is similar and the user confirms that the semantics are the same, storing the organization name in the organization name set that the user confirms that the semantics are the same can realize the self-learning of the voice robot, that is, the next time a query for the organization name appears, it can be directly located in the corresponding organization name set.

[0061] Please refer to Figure 6 , which shows a flowchart of another voice number query method provided by an embodiment of the present invention, the flowchart is mainly Figure 4 A flowchart of the steps further defined in step 401 of the process "If there is only one second institution name in the set of multiple institution names that is similar to the first institution name, confirm with the user whether the first institution name and the second institution name are semantically identical."

[0062] like Figure 6 As shown, in step 601, the location of the number of the user's incoming call is queried;

[0063] In step 602, if there are multiple second institution names in the set of multiple institution names that are similar to the first institution name, it is determined whether there is an institution name related to the place where the number is located in the multiple second institution names;

[0064] In step 603, if there is a third institution name related to the number's location among the multiple second institution names, confirm with the user whether the third institution name is semantically identical to the second institution name.

[0065] In this embodiment, for step 601, the voice robot queries the location of the number of the user's incoming call. For example, the voice robot determines the location of the user's number and the specific city based on the user's location information and phone number information.

[0066] Then for step 602, the voice robot searches for multiple second institution names in the set of multiple institution names that are similar to the first institution name, and determines whether there is an institution name related to the number's location among the multiple second institution names. For example, the institution name that the user needs to find is "Jiaotong University", and the voice robot searches for "Shanghai Jiaotong University", "Beijing Jiaotong University" and "Nanjing Jiaotong University" in the set of multiple institution names. The voice robot determines that the user's location is "Shanghai" based on the user's number's location or the user's location information. The voice robot searches for the multiple institution name sets "Shanghai Jiaotong University", "Beijing Jiaotong University" and "Nanjing Jiaotong University" based on the user's location information to determine whether there is an institution name similar to the user's area.

[0067] Finally, for step 603, the voice robot finds out that among the multiple second institution names, there is a third institution name related to the place where the number is located, and confirms with the user whether the third institution name is semantically identical to the second institution name. For example, the voice robot finds "Shanghai Jiao Tong University", "Shanghai Public Transport Group" and "Shanghai Traffic Police Detachment" based on the user's location information, and confirms with the user whether there is the institution that needs to be found.

[0068] The method of this embodiment improves the efficiency of searching for the name of the organization that the user needs to find by determining the location of the user number, which also makes the voice robot more accurate in the search process.

[0069] It should be noted that the above method steps are not used to limit the execution order of each step. In fact, some steps may be executed simultaneously or in the opposite order to the specified step order, and this application has no limitation on this.

[0070] The following describes some problems encountered by the inventor in the process of implementing the present invention and a specific embodiment of the solution finally determined, so that those skilled in the art can better understand the solution of the present application.

[0071] In the process of implementing the present application, the inventors found that the traditional directory query scheme has the following disadvantages:

[0072] (1) When users search for numbers through online search engines, the returned results cover a wide range of areas, making it difficult to quickly locate valuable information and resulting in low search efficiency. For questions and answers on forums, this is often not effective. The results of a user's search may be records from three to five years ago, and the actual organization's phone number may have been updated; for the official website of an institution or organization, users often need to find the phone number of a certain department, and they need to first find the entrance to the corresponding department on the official website and then enter to find the department's phone number.

[0073] (2) The method of checking the number through the hotline 114 and finding the phone number after the manual agent searches the directory is inefficient. After receiving the user's call, the 114 service personnel need to manually query the system and then return the result to the user by phone. This series of actions are all completed by manual agents, which is time-consuming and inefficient. One agent can only connect to one user at a time. In addition, manual agents are limited by time. If they want to provide 24-hour service, they need to work in shifts, which is costly.

[0074] The existing technical solutions have the following defects: they cannot flexibly handle various abbreviations of the same organization, and they do not have self-learning capabilities, which will greatly reduce the experience of voice interaction with users.

[0075] 1) Because institutions have different abbreviations, for example, the abbreviations of "University of Science and Technology of China" include "USTC", "USTC", "China University of Science and Technology" and many other abbreviations. These two methods are based on the number repository to query numbers, and cannot generalize abbreviations and terms, so the effect is not good;

[0076] 2) Due to the spoken language, users have many different ways of saying institutions, such as "USTC in Hefei", "USTC in Hefei", etc. The system cannot generalize the abbreviations based on user statements through self-learning.

[0077] The inventor found in the process of implementing this application that it is not easy for those skilled in the art to think of the technical solution of this application because, on the one hand, 114 directory enquiry scenarios are mostly provided by human agents, and the same organization has different titles, and voice robots cannot flexibly handle abbreviations, which increases the difficulty of digitalization and intelligentization of directory enquiry services; on the other hand, AISpeech has a leading advantage in voice interaction experience for many years, and has introduced voice robots required by various scenarios for government affairs construction, with several landing cases. These lay a solid foundation for the conception and implementation of this patent.

[0078] Please refer to Figure 7 , which shows the storage structure of the directory query scheme provided by an embodiment of the present application.

[0079] In the present invention, we propose a self-learning optimization method and system for voice number query. Figure 7 The storage structure of the solution proposed in this patent is shown. It not only has a simple "institution name-telephone" correspondence, but also includes a set of abbreviations. When the institution name spoken by the user hits the full name of the institution or its corresponding abbreviation set, the system broadcasts the telephone number through voice. Taking "Beijing University of Posts and Telecommunications" as an example, modern people are more accustomed to calling it by its abbreviation: Beiyou; but people who lived in the 1970s and 1980s are accustomed to calling it "Beijing University of Posts and Telecommunications." Both names refer to the same institution, so a set of abbreviations is added to cope with flexible conversation scenarios. Compared with the traditional 114 directory query, this solution not only breaks through the limitations of manual services, but also realizes the query of the abbreviation set of the institution on the basis of querying the full name, effectively dealing with different abbreviations of the same institution in actual conversations.

[0080] Please refer to Figure 8 , which shows a solution architecture diagram provided by an embodiment of the present application.

[0081] The relevant service modules of this solution are: ASR (Automatic Speech Recognition), NLU (Natural Language Understanding), NLG (Natural Language Generation), TTS (Text To Speech), SSM (Semantic Similarity Matching); the protocol for accessing the dialogue center is SIP (Session Initiation Protocol). Figure 8The architecture diagram of the voice directory query self-learning optimization method is shown. When the organization name ${query} obtained through voice recognition does not match the full name nor the abbreviation set, the system matches ${query} with the existing full name and abbreviation set through semantic similarity, selects the organization full name ${orgname} with the highest similarity, and asks the user whether it is this organization; if the system is confirmed, it is added to the abbreviation set of this organization, thus realizing dialogue self-learning. When the next user asks for the phone number of this organization with the same abbreviation, the system can find the name in the abbreviation set because it has already learned it.

[0082] Please refer to Fig. 9 , which shows an example of an actual conversation of the voice directory inquiry solution provided by an embodiment of the present application.

[0083] Fig. 9 List the possible dialogue processes when actual users interact with the system: Assume that in the system's data storage, there is no element "Beijing University of Posts and Telecommunications" in the abbreviation set corresponding to the institution "Beijing University of Posts and Telecommunications". User 1 asks for the phone number of Beijing University of Posts and Telecommunications, and the system records the institution ${query} = "Beijing University of Posts and Telecommunications". The search for "Beijing University of Posts and Telecommunications" in the full name and abbreviation set is fruitless. After similarity comparison between ${query} and the full name and abbreviation set, "University of Posts and Telecommunications" scores the highest. The full name ${orgname} corresponding to the set where the element "University of Posts and Telecommunications" is located is "Beijing University of Posts and Telecommunications". The voice robot will ask the user whether he refers to "Beijing University of Posts and Telecommunications". After receiving an affirmative answer from the user, the number ${tel} = 11**** corresponding to "Beijing University of Posts and Telecommunications" will be broadcast to the user, and the title "Beijing University of Posts and Telecommunications" will be added to the abbreviation set. Thereafter, if user 2 inquires about the phone number of "Beijing University of Posts and Telecommunications" and asks for "Beijing University of Posts and Telecommunications", then the system that has been self-learning will query the updated abbreviation set (refer to the storage structure). Figure 7 ), will directly return to the phone number of the user "Beijing University of Posts and Telecommunications".

[0084] Please refer to Fig.10 , which shows a software implementation flow chart provided by an embodiment of the present application.

[0085] Fig.10The software implementation flow chart of the directory query solution proposed in this patent is shown: after the user answers with voice, the system converts the voice into text through ASR, and then understands the user's directory query intention through NLU and extracts the name of the organization entity in the text. If the name is found in the set of full names and abbreviations of the organization, the phone number will be broadcast to the user; if the query in the database fails, the system will match the name with the set of full names and abbreviations of the organization, obtain the full name of the organization ${orgname} with a high similarity score, and then confirm with the user whether it is this organization. After understanding the user's confirmation intention, the corresponding phone number will be broadcast to the user, and then the new organization name will be added to the phone directory, so as to achieve conversational self-study.

[0086] After the user dials the hotline, the voice robot extracts the name of the organization the user wants to query through voice interaction with the user. If the organization name is already in the full name or the corresponding abbreviation set, the robot will broadcast the corresponding phone number to the user; if the organization is no longer in the directory, the directory inquiry system will compare the user's name for the organization with the organization in the directory through semantic similarity. If the similarity calculation result is greater than a certain set threshold, the user will be asked to confirm whether it is an organization already in the directory; if the user gives a positive answer, the system will add the new name to the database, and continuously update it so that other users can query the results faster when they ask. Through a series of operations such as query-similarity matching-broadcast and update, the system has a good closed loop and can adapt to the diversity of language expression.

[0087] The solution of this application is designed and optimized mainly from the following aspects:

[0088] The directory inquiry systems currently on the market can neither flexibly handle various abbreviations of the same organization nor have self-learning capabilities, which will greatly reduce the user's voice interaction experience. This invention mainly improves and optimizes these aspects to enhance the user's experience and practicality of voice interaction.

[0089] Please refer to Fig.11 , which shows a block diagram of a voice directory inquiry device provided by an embodiment of the present invention.

[0090] like Fig.11 As shown, the voice directory query device 1100 includes an extraction program module 1110 , an existence processing program module 1120 , a similarity determination program module 1130 , a confirmation program module 1140 and a feedback program module 1150 .

[0091] Among them, the extraction program module 1110 is configured to extract the first organization name from the voice directory information in response to the voice robot receiving the voice directory information input by the user, and query whether the first organization name exists in the preset multiple organization name sets; the existence processing program module 1120 is configured to feedback the number information corresponding to the first organization name to the user if the first organization name exists in the multiple organization name sets; the similarity judgment program module 1130 is configured to, if the first organization name does not exist in the preset multiple organization name sets, the voice robot determines whether there is a second organization name similar to the first organization name in the multiple organization name sets; the confirmation program module 1140 is configured to, if there is a second organization name similar to the first organization name in the multiple organization name sets, confirm to the user whether the first organization name and the second organization name are semantically identical; and the feedback program module 1150 is configured to, if the user's semantic identity instruction is received, the voice robot feedbacks the number information corresponding to the second organization name to the user.

[0092] It should be understood that Fig.11 Modules and references documented in Figure 1 , Figure 2 , Figure 3 , Figure 4 Figure 5 and Figure 6 Therefore, the operations and features described above for the method and the corresponding technical effects are also applicable to Fig.11 The modules in it will not be described in detail here.

[0093] It is worth noting that the modules in the embodiments of the present disclosure are not used to limit the solutions of the present disclosure. For example, the judgment module can be described as a module that judges whether the interactive state is a playback scene when the device is in an interactive state. In addition, the relevant functional modules can also be implemented by a hardware processor, for example, the judgment module can also be implemented by a processor, which will not be repeated here.

[0094] An embodiment of the present invention further provides a computer program product, which includes a computer program stored on a non-volatile computer-readable storage medium. The computer program includes program instructions. When the program instructions are executed by a computer, the computer executes any of the above-mentioned voice directory query methods.

[0095] Fig.12 is a schematic diagram of the structure of an electronic device provided by an embodiment of the present invention, such as Fig.12 As shown, the device includes: one or more processors 1201 and a memory 1202, Fig.12A processor 1201 is used as an example. The device of the voice number query method may also include: an input device 1203 and an output device 1204. The processor 1201, the memory 1202, the input device 1203 and the output device 1204 may be connected via a bus or other means. Fig.12 In the example, the bus connection is used. The memory 1202 is the above-mentioned non-volatile computer-readable storage medium. The processor 1201 executes various functional applications and data processing of the server by running the non-volatile software programs, instructions and modules stored in the memory 1202, that is, the voice directory inquiry method of the above-mentioned method embodiment is realized. The input device 1203 can receive input digital or character information, and generate key signal input related to the user setting and function control of the voice directory inquiry device. The output device 1204 may include a display device such as a display screen.

[0096] The above product can execute the method provided by the embodiment of the present invention, and has the functional modules and beneficial effects corresponding to the execution method. For technical details not described in detail in this embodiment, please refer to the method provided by the embodiment of the present invention.

[0097] As an implementation mode, the electronic device is applied to a voice directory inquiry device and is used for a client, and includes: at least one processor; and a memory connected to the at least one processor in communication; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can:

[0098] In response to the voice robot receiving the voice directory inquiry information input by the user, extracting the first organization name from the voice directory inquiry information, and querying whether the first organization name exists in a preset set of multiple organization names;

[0099] If the first institution name exists in the set of multiple institution names, feeding back number information corresponding to the first institution name to the user;

[0100] If the first institution name does not exist in the preset multiple institution name sets, the voice robot determines whether there is a second institution name similar to the first institution name in the multiple institution name sets;

[0101] If there is a second institution name similar to the first institution name in the set of multiple institution names, confirming with the user whether the first institution name and the second institution name are semantically identical;

[0102] If receiving a semantically identical instruction from the user, the voice robot will feed back the number information corresponding to the second institution name to the user.

[0103] The electronic device of the embodiment of the present application exists in various forms, including but not limited to:

[0104] (1) Mobile communication devices: These devices are characterized by their mobile communication functions and their main purpose is to provide voice and data communications. These terminals include: smart phones (such as iPhone), multimedia phones, functional phones, and low-end phones.

[0105] (2) Ultra-mobile personal computer devices: These devices fall into the category of personal computers, have computing and processing capabilities, and generally also have mobile Internet access features. These terminals include: PDA, MID and UMPC devices, such as iPad.

[0106] (3) Portable entertainment devices: These devices can display and play multimedia content. They include audio and video players (such as iPods), handheld game consoles, e-books, as well as smart toys and portable car navigation devices.

[0107] (4) Server: A device that provides computing services. The server consists of a processor, hard disk, memory, system bus, etc. The server is similar to a general computer architecture, but because it needs to provide highly reliable services, it has higher requirements in terms of processing power, stability, reliability, security, scalability, and manageability.

[0108] (5) Other electronic devices with data interaction functions.

[0109] The device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, i.e., they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Those of ordinary skill in the art may understand and implement it without creative effort.

[0110] Through the description of the above implementation methods, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a necessary general hardware platform, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solution is essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a disk, an optical disk, etc., including 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 of each embodiment or some parts of the embodiment.

[0111] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A voice number query method, comprising: In response to the voice robot receiving the voice directory inquiry information input by the user, extracting the first organization name from the voice directory inquiry information, and querying whether the first organization name exists in a preset plurality of organization name sets, wherein each of the plurality of organization name sets has a complete organization name uniquely corresponding to the organization name set, and the organization name set includes the abbreviation of the complete organization name and the full name of the complete organization name; If the first institution name exists in the set of multiple institution names, feeding back number information corresponding to the first institution name to the user; If the first institution name does not exist in the preset multiple institution name sets, the voice robot determines whether there is a second institution name similar to the first institution name in the multiple institution name sets; If there is a second institution name similar to the first institution name in the set of multiple institution names, confirming with the user whether the first institution name and the second institution name are semantically identical; If there is only one second institution name in the set of multiple institution names that is similar to the first institution name, confirming with the user whether the first institution name and the second institution name are semantically identical; If there are multiple second institution names in the set of multiple institution names that are similar to the first institution name, confirming with the user whether any second institution name in the multiple second institution names is semantically identical; If receiving a semantically identical instruction from the user, the voice robot feeds back the number information corresponding to the name of the second organization to the user; If there is only one second organization name that is similar to the first organization name, and the user confirms that the semantics are the same, the first organization name is stored in the organization name set corresponding to the second organization name to enrich the organization name set to achieve self-learning of the voice robot, and the number information corresponding to the second organization name is fed back to the user; If receiving a semantically identical instruction from the user, the voice robot will feed back the number information corresponding to the second institution name to the user.

2. The method according to claim 1, wherein: The voice robot is preset with a semantic similarity threshold, and the determining whether there is a second institution name similar to the first institution name in the set of multiple institution names includes: Calculating the semantic similarity between the abbreviations in the plurality of institution name sets and the first institution name; If there is a second institution name in at least one institution name set whose similarity to the first institution name is greater than or equal to the semantic similarity threshold, it is confirmed that there is a second institution name similar to the first institution name in the multiple institution name sets.

3. The method according to claim 2, wherein: The method further comprises: If there are multiple second organization names whose similarity to the first organization name is greater than or equal to the semantic similarity threshold, only the second organization name with the highest similarity is obtained.

4. The method according to claim 1, wherein: The voice robot feeding back the number information corresponding to the second organization name to the user includes: Obtain the complete institution name of the institution name set corresponding to the second institution name; The number information corresponding to the second institution name and the corresponding complete institution name are fed back to the user.

5. The method according to claim 1, wherein: There are multiple second institution names in the set of multiple institution names that are similar to the first institution name, and confirming to the user whether any second institution name in the multiple second institution names is semantically identical comprises: Check the location of the user's incoming call number; If there are multiple second institution names in the set of multiple institution names that are similar to the first institution name, determining whether there is an institution name related to the place where the number is located in the multiple second institution names; If there is a third institution name related to the number's location among the multiple second institution names, confirm with the user whether the third institution name is semantically identical to the second institution name.

6. A voice directory inquiry device, comprising: An extraction program module is configured to extract a first organization name from the voice directory inquiry information in response to the voice robot receiving the voice directory inquiry information input by the user, and query whether the first organization name exists in a preset set of multiple organization names; an existence processing program module, configured to feed back number information corresponding to the first institution name to the user if the first institution name exists in the set of multiple institution names; a similarity determination program module, configured to, if the first institution name does not exist in the preset multiple institution name sets, the voice robot determines whether there is a second institution name similar to the first institution name in the multiple institution name sets; a confirmation program module, configured to confirm with the user whether the first institution name is semantically identical to the second institution name if there is a second institution name similar to the first institution name in the set of multiple institution names; The feedback program module is configured such that if the voice robot receives a semantically identical instruction from the user, it will feed back the number information corresponding to the name of the second organization to the user.

7. An electronic device, comprising: At least one processor, and a memory communicatively connected to the at least one processor, wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the steps of the method described in any one of claims 1 to 5.

Images