Intelligent automatic question-answering method, system, server and storage medium

Abstract

The present invention discloses a method and system for intelligent automatic question answering, the method comprising: receiving initial request information of a user, the initial request information comprising a first instance and a third instance; searching a knowledge point corresponding to the initial request information in a knowledge base through similarity calculation; when the corresponding knowledge point is not found in the knowledge base, performing reasoning processing according to a pre-established knowledge graph, the reasoning processing comprising: querying a first instance in a pre-established knowledge graph; obtaining an association path of the first instance in the knowledge graph until the association path comprises a third instance; determining an answer according to the first instance, the third instance and the association path, the association path comprising M attribute information, M being greater than or equal to 2; and sending the determined answer to the user. The present invention can greatly improve the accuracy of intelligent automatic question answering.

Description

Technical Field

[0001] The embodiments of the present invention relate to human-computer interaction technology, and more particularly to an intelligent automatic question-answering method, an intelligent automatic question-answering system, a server and a storage medium. Background Art

[0002] Human-computer interaction is a science that studies the interactive relationship between systems and users. Systems can be various machines or computerized systems and software. Intelligent automatic question-answering systems are an artificial intelligence system developed based on human-computer interaction technology, such as intelligent customer service systems, voice control systems, etc.

[0003] The knowledge base is the foundation of the intelligent automatic question-answering system. In the knowledge base, information is effectively organized for retrieval and utilization. The knowledge base stores multiple knowledge points, each of which includes one or more questions and corresponding answers. When the user enters the request information, the semantic similarity between the request information and the questions in the knowledge base is calculated. If there are questions with a similarity greater than a first preset threshold, the answer corresponding to the question with the greatest similarity is returned to the user.

[0004] Generally, the knowledge base stores relatively simple knowledge points, such as:

[0005] Knowledge point 1: Who is A's ex-husband?

[0006] Knowledge point 2: Who is B's current girlfriend?

[0007] Knowledge point 3: Who was C's first husband?

[0008] Knowledge point 4: Who is D's eldest daughter?

[0009] When the request information is: what is the relationship between A and D, what is the relationship between A and E, or what is the relationship between C and E, etc., according to the existing question-answering technology, the corresponding answer cannot be obtained in the knowledge base at this time. Summary of the invention

[0010] In order to solve the above technical problems, an embodiment of the present invention provides a method for intelligent automatic question answering, a system, a server and a storage medium for intelligent automatic question answering, so as to find answers to complex questions without changing the knowledge points stored in the knowledge base, thereby improving the accuracy of question answering.

[0011] The embodiment of the present invention provides a method for intelligent automatic question answering, including:

[0012] Receiving initial request information from a user, wherein the initial request information includes a first instance and a third instance;

[0013] Searching for a knowledge point corresponding to the initial request information in the knowledge base through similarity calculation;

[0014] When the corresponding knowledge point is not found in the knowledge base, reasoning processing is performed according to the pre-established knowledge graph, and the reasoning processing includes:

[0015] Querying a first instance in a pre-established knowledge graph;

[0016] Obtaining an association path of the first instance in the knowledge graph until the association path includes a third instance;

[0017] Determine an answer according to the first instance, the third instance and the association path, wherein the association path includes M attribute information, where M is greater than or equal to 2;

[0018] Send the confirmed answer to the user.

[0019] Optionally, the knowledge graph records the association relationship between multiple instances in a table.

[0020] Optionally, the initial request information includes: what is the relationship between the first instance and the third instance or what is the relationship between the first instance and the third instance.

[0021] Optionally, searching the knowledge base for knowledge points corresponding to the initial request information through similarity calculation includes: calculating the semantic similarity between the initial request information and the questions stored in the knowledge base, comparing the calculated maximum similarity with a preset first preset threshold, and when the calculated similarity is greater than the first preset threshold, using the knowledge point corresponding to the similarity as the knowledge point to be searched; otherwise, the corresponding knowledge point is not found in the knowledge base.

[0022] The embodiment of the present invention further provides an intelligent automatic question-answering system, comprising: a user request receiving module, configured to receive initial request information of a user, wherein the initial request information includes a first instance and a third instance;

[0023] A knowledge point search module, used to search the knowledge point corresponding to the initial request information in the knowledge base through similarity calculation;

[0024] The reasoning module is used to perform reasoning processing according to the pre-established knowledge graph when the corresponding knowledge point is not found in the knowledge base. The reasoning module includes:

[0025] An instance query unit, configured to query a first instance in a pre-established knowledge graph;

[0026] an association path acquisition unit, configured to acquire an association path of the first instance in the knowledge graph until the association path includes a third instance;

[0027] an answer determination unit, configured to determine an answer according to the first instance, the third instance, and the association path, wherein the association path includes M attribute information, where M is greater than or equal to 2;

[0028] The output module is used to send the determined answer to the user.

[0029] Optionally, the knowledge graph records the association relationship between multiple instances in a table.

[0030] Optionally, the initial request information includes: what is the relationship between the first instance and the third instance or what is the relationship between the first instance and the third instance.

[0031] Optionally, the knowledge point search module calculates the semantic similarity between the initial request information and the questions stored in the knowledge base, compares the calculated maximum similarity with a preset first preset threshold, and when the calculated similarity is greater than the first preset threshold, the knowledge point corresponding to the similarity is used as the knowledge point to be searched; otherwise, the corresponding knowledge point is not found in the knowledge base.

[0032] The embodiment of the present invention further provides a server, including:

[0033] one or more processors;

[0034] a storage device for storing one or more programs,

[0035] When the one or more programs are executed by the one or more processors, the one or more processors implement the above-mentioned intelligent automatic question-answering method.

[0036] An embodiment of the present invention further provides a computer storage medium on which a computer program is stored, and when the program is executed by a processor, the above-mentioned intelligent automatic question-answering method is implemented.

[0037] The technical solution of the embodiment of the present invention performs inference processing to obtain inferred request information when the knowledge point corresponding to the initial request information cannot be found in the knowledge base, and searches the knowledge base for the knowledge point corresponding to the inferred request information through similarity calculation. When the corresponding knowledge point is not found in the knowledge base, the above-mentioned reasoning and calculation process is repeated until the corresponding knowledge point is found in the knowledge base, thereby achieving the goal of finding answers to complex questions without changing the knowledge points stored in the knowledge base, and improving the accuracy of question and answer and saving storage space without changing the content stored in the knowledge base. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] Figure 1is a flow chart of an intelligent automatic question-answering method provided by Embodiment 1 of the present invention;

[0039] Figure 2 is a structural diagram of an intelligent automatic question-answering system provided by Embodiment 2 of the present invention;

[0040] Figure 3 A schematic diagram of the structure of a server provided in Embodiment 3 of the present invention. DETAILED DESCRIPTION

[0041] The present invention will be further described in detail below in conjunction with the accompanying drawings and embodiments. It is to be understood that the specific embodiments described herein are only used to explain the present invention, rather than to limit the present invention. It should also be noted that, for ease of description, only the parts related to the present invention are shown in the accompanying drawings, rather than all of the contents.

[0042] In order to facilitate understanding of the solution of the embodiment of the present invention, the basic knowledge of the intelligent automatic question answering system is briefly introduced:

[0043] 1. Knowledge points

[0044] The most primitive and simplest form of basic knowledge points in the knowledge base is the FAQ commonly used in daily life, which is generally in the form of a "question-answer" pair. For example, "the price of ringback tone" is a standard question description that expresses clearly. The "question" here should not be understood in a narrow sense as "inquiry", but should be understood in a broad sense as an "input" with a corresponding "output". For example, for semantic recognition used for control systems, a user's instruction, such as "turn on the radio", should also be understood as a "question", and the corresponding "answer" can be the call of a control program used to execute the corresponding control.

[0045] When users input into the machine, the ideal situation is to use standard questions, so that the machine's intelligent semantic recognition system can immediately understand what the user means. However, users often do not use standard questions, but some variant forms of standard questions. For example, if the standard form of the question for switching radio stations is "change a station", then the user may use the command "switch a station", and the machine also needs to be able to recognize that the user is expressing the same meaning.

[0046] Therefore, for intelligent semantic recognition, the knowledge base needs to have extended questions of standard questions, which have slightly different expressions from standard questions but express the same meaning.

[0047] Therefore, the knowledge base includes multiple knowledge points, each knowledge point includes questions and answers, and the questions include standard questions and multiple extended questions.

[0048] The questions in the knowledge points are generally expressed in the form of semantic expressions. The following are the symbols in the semantic expressions:

[0049] 1.1.1 Representation of word classes ([])

[0050] To distinguish between words and word classes in expressions, word classes must appear in square brackets "[]". Word classes that appear in square brackets are generally "narrow word classes", but "broad word classes" can also be supported by configuring system parameters. The following are some examples of simple expressions: [Fetion][how][open], [introduction][MMS][service], [Fetion]'s [login][method], [call reminder][how][charge].

[0051] 1.1.2 Expression of OR relationship (|)

[0052] The word classes in square brackets can appear multiple times through the "or" relationship. These word classes in the "or" relationship will be calculated separately in the "expanded" way when calculating the similarity. "Expanding" is mainly the process of expanding the semantic expression into multiple simple expressions based on the meaning of "or". For example: [color ringtone] [activation] [method | steps] can be expanded into "[color ringtone] [activation] [steps]" and "[color ringtone] [activation] [method]" two simple semantic expressions. Examples of this type of semantic expression are as follows: [how to] [query | know] [PUK code], [unsubscribe | cancel | close | deactivate] [IP | 17951] [domestic long-distance discount package].

[0053] 1.1.3 Non-essential representation (?)

[0054] The word class in square brackets can be added with "?" at the end to indicate that it may or may not appear, that is, a non-essential relationship. This non-essential word class will also be calculated separately in the "expanded" way when calculating similarity. "Expanding" is mainly the process of expanding the semantic expression containing non-essential word classes (or "or combinations" of word classes) into two simple semantic expressions containing and not containing this word class.

[0055] For example, [introduction][mobile video][military column][content][what?] can be expanded into two simple semantic expressions: “[introduction][mobile video][military column][content]” and “[introduction][mobile video][military column][content][what]”.

[0056] 2. Ontology

[0057] Ontology is a formal expression of a set of concepts in a specific field and their relationships. It is an abstraction of a field. We use ontology to obtain knowledge in a certain field. Ontology describes the concepts in the field and the relationships between these concepts.

[0058] 2.1 Ontology composition

[0059] The ontology consists of ontology classes and instances. The ontology class defines the structure of the corresponding instance, and the instance is an individual object that conforms to the ontology class definition.

[0060] 2.1.1 Ontology Class

[0061] In the real world, there are often objects belonging to the same class (e.g., a car is just one of many cars in the world), and there are also many different objects that share the same characteristics. The common characteristics of these objects can be used to establish a prototype for them. The ontology class is a definition of this prototype. It is an abstraction of all individuals that have or share common characteristics. The characteristics of the ontology class are called class properties. For example, the ontology class "person" has properties such as "name", "age" and "weight". Class properties consist of attribute names and a set of semantic expressions. In the ontology class, we can use a set of semantic expressions to describe the semantics expressed by the attributes. The semantic expression of the class attribute consists of instance symbols, word classes, and grammatical symbols. The instance symbol is used to represent the word class or semantic block that describes the individual.

[0062] 2.1.2 Instance

[0063] An instance is the product of the instantiation of an ontology class, representing a specific individual object that conforms to the definition of the ontology class. An instance consists of an instance name, instance semantics, attributes, and attribute values. Instance semantics: Instance semantics is a semantic expression used to describe instance semantics. This semantic expression replaces the "instance symbol" in the semantic expression of the ontology class attribute during the instantiation of the ontology class, thereby generating the semantics of the instance attribute.

[0064] Instance attributes: Instance attributes include general attributes and custom attributes. General attributes are attributes defined in the ontology class of the instantiated instance. For example, the "Santana Sedan" instance is instantiated from the "Sedan" ontology class, so the "wheelbase", "weight" and "displacement" attributes defined in the "Sedan" ontology class are general attributes. Custom attributes are customized feature descriptions that are only suitable for the individual and are customized outside of general attributes in order to more comprehensively describe the individual.

[0065] Attribute value: Attribute value is the specific description of the attribute, which is used to accurately express the quality or quantity of the attribute of the instance. For example, "male" is the attribute value of the "gender" attribute in the instance "Zhang San".

[0066] Embodiment 1

[0067] Figure 1 : is a flowchart of an intelligent automatic question-answering method provided by Embodiment 1 of the present invention. This embodiment is applicable to the case of searching for knowledge points corresponding to user request information including two instances. The method can be executed by an intelligent automatic question-answering system, which can be implemented by software and / or hardware and can generally be integrated on a server or other device. The method specifically includes the following steps:

[0068] Step 210: receiving initial request information from a user, wherein the initial request information includes a first instance and a third instance.

[0069] The user's initial request information includes two instances. For example, the initial request information is "the relationship between A and B", wherein A is the first instance and B is the third instance; or A is the third instance and B is the first instance.

[0070] Specifically, the initial request information may include: what is the relationship between the first instance and the third instance or what is the relationship between the first instance and the third instance, etc.

[0071] Step 220: searching the knowledge base for knowledge points corresponding to the initial request information by similarity calculation.

[0072] The semantic similarity between the initial request information and the questions stored in the knowledge base is calculated, and the calculated maximum similarity is compared with a preset first preset threshold to find the knowledge point corresponding to the initial request information.

[0073] Specifically, searching the knowledge point corresponding to the initial request information in the knowledge base through similarity calculation includes: calculating the semantic similarity between the initial request information and the question stored in the knowledge base, comparing the calculated maximum similarity with a preset first preset threshold, and when the calculated similarity is greater than the first preset threshold, using the knowledge point corresponding to the similarity as the knowledge point to be searched; otherwise, the corresponding knowledge point is not found in the knowledge base.

[0074] Step 230, when the corresponding knowledge point is not found in the knowledge base, reasoning processing is performed according to the pre-established knowledge graph.

[0075] When the calculated similarities are all less than the first preset threshold, it is determined that the knowledge base does not directly store the knowledge point corresponding to the initial request information, and then reasoning processing can be performed based on the pre-established knowledge graph. The knowledge graph includes multiple instances in the knowledge base and the associations between the instances.

[0076] The inference process includes:

[0077] Step 231, query the first instance in the pre-established knowledge graph.

[0078] Specifically, the knowledge graph may be a table that records the association relationship between multiple instances. When performing reasoning, the first instance is first searched to obtain the association path between the first instance and the third instance.

[0079] Step 232, obtaining the association path of the first instance in the knowledge graph until the association path includes the third instance.

[0080] Among them, the association path is the path of multiple instances in the association knowledge graph, including at least two instances, and knowledge points that respectively include at least two instances, a knowledge point includes a question and an answer, wherein the question includes an instance, and the answer includes another instance, and the knowledge points that respectively include one instance in the question and the answer are the association paths of the instances respectively included in the question and the answer. For example, the knowledge points corresponding to instance 1 include knowledge points 1.1, 1.2, etc., the answer to knowledge point 1.1 refers to instance 2, the knowledge points corresponding to instance 2 include knowledge points 2.1, 2.2, etc., and the answer to knowledge point 2.2 refers to instance 3, then the association relationship between instance 1 and instance 2 is the association path of instance 1 and instance 2, and the association path includes instance 1, knowledge point 1.1 and instance 2, and the association relationship between instance 1, instance 2 and instance 3 is the association path of instance 1 and instance 3, that is, the association path includes instance 1, knowledge point 1.1, instance 2, knowledge point 2.2 and instance 3.

[0081] Step 233: determine an answer based on the first instance, the third instance and the associated path.

[0082] In the knowledge base of this embodiment, each knowledge point only stores one attribute information between two instances, and the association path includes M attribute information, where M is greater than or equal to 2, thereby making up for the defect that the existing knowledge base only stores simple knowledge points.

[0083] An answer corresponding to the user's initial request information can be determined according to the first instance, the third instance and the associated path, and the answer is returned to the user.

[0084] In this embodiment, the text answer can be directly returned to the user, or the text answer can be converted into a voice format and then sent to the user, which does not limit the protection scope of the present invention.

[0085] For example, the user's initial request information is "What is the relationship between Huawei P9 and Mr. Ren Zhengfei?" The initial request information includes the first instance "Huawei P9" and the third instance "Mr. Ren Zhengfei". The semantic similarity between the initial request information and the question in the knowledge base is calculated. The calculated similarities are all less than the first preset threshold, indicating that the knowledge point corresponding to the initial request information cannot be found in the knowledge base. At this time, the reasoning process based on the knowledge graph is entered: the first instance "Huawei P9" is queried in the pre-established knowledge graph, and the association path of the first instance "Huawei P9" in the knowledge graph is obtained. The process of obtaining the association path is: The knowledge points corresponding to the first instance "Huawei P9" are knowledge points 1.1 and 1.2, among which the question of knowledge point 1.1 is "What is the CPU model of Huawei P9?" The answer is to quote the second instance "HiSilicon Kirin 955". Then the knowledge point corresponding to the second instance "HiSilicon Kirin 955" is searched, and knowledge point 2.1 is found. The question is "Who is the manufacturer of HiSilicon Kirin 955?" The answer is to quote the fourth instance "HiSilicon Semiconductor Co., Ltd." The knowledge points corresponding to the fourth instance do not reference other instances, so no association path including the third instance "Mr. Ren Zhengfei" is found; the question of knowledge point 1.2 is "Who is the manufacturer of Huawei P9?", and the answer is to reference the fifth instance "Huawei Technologies Co., Ltd.". The knowledge point corresponding to the fifth instance "Huawei Technologies Co., Ltd." is knowledge point 5.1, and the question of knowledge point 5.1 is "Who is the president of Huawei Technologies Co., Ltd.", and the answer is to reference the third instance "Mr. Ren Zhengfei". The association path includes the third instance "Mr. Ren Zhengfei", that is, the association path including the third instance "Mr. Ren Zhengfei" includes the first instance, knowledge point 1.2, the fifth instance, knowledge point 5.1 and the third instance; therefore, the answer is determined to be "Mr. Ren Zhengfei is the president of the manufacturer of Huawei P9" based on the first instance, the third instance and the association path, and the answer is returned to the user.

[0086] For another example, the user's initial request information is "What is the relationship between A and B?" The initial request information includes the first instance "A" and the third instance "B". The semantic similarity between the initial request information and the question in the knowledge base is calculated. The calculated similarities are all less than the first preset threshold, indicating that the knowledge point corresponding to the initial request information cannot be found in the knowledge base. At this time, the reasoning process based on the knowledge graph is entered: the first instance "A" is queried in the pre-established knowledge graph, and the association path of the first instance "A" in the knowledge graph is obtained. The process of obtaining the association path is: the knowledge points corresponding to the first instance "A" are 1.1, The question of knowledge point 1.1 is "Who is A's ex-husband?", and the corresponding answer is to quote the second instance "C". Then, it is found that the knowledge point corresponding to the second instance "C" is 2.1. The question of knowledge point 2.1 is "Who is C's current girlfriend?", and the corresponding answer is to quote the fourth instance "D". Then, it is found that the knowledge point corresponding to the fourth instance "D" is 4.1. The question of knowledge point 4.1 is "Who is D's daughter?", and the answer is to quote the third instance "B". The association path including the third instance "B" is found, that is, the association path including the third instance "B" includes the first instance "B", knowledge point 1.1, the second instance "C", knowledge point 2.1, the fourth instance "D", knowledge point 4.1 and the third instance "B". Therefore, according to the first instance, the third instance and the association path, the answer is determined to be "B is the daughter of A's ex-husband's current girlfriend", and the answer is returned to the user.

[0087] The technical solution of this embodiment is to perform reasoning processing according to a pre-established knowledge graph when the user's initial request information includes the first instance and the third instance, and the knowledge point corresponding to the initial request information cannot be found in the knowledge base through similarity calculation. The reasoning processing includes: querying the first instance in the pre-established knowledge graph, obtaining the association path of the first instance in the knowledge graph, until the association path includes the third instance, and determining the answer according to the first instance, the third instance and the association path, thereby achieving the goal of finding the answer corresponding to the request information including two instances without changing the knowledge points stored in the knowledge base, and improving the accuracy of question and answer and saving storage space without changing the content stored in the knowledge base.

[0088] Embodiment 2

[0089] Figure 2 It is a structural diagram of an intelligent automatic question-answering system provided in the second embodiment of the present invention. This embodiment can be applied to the situation of searching for knowledge points corresponding to user request information including two instances. The system can be implemented by software and / or hardware and can generally be integrated on a server and other equipment. The device includes: a user request receiving module 410, a knowledge point search module 420, an inference module 430 and an output module (not shown in the figure).

[0090] The user request receiving module 410 is used to receive the initial request information of the user, wherein the initial request information includes the first instance and the third instance;

[0091] A knowledge point search module 420, configured to search the knowledge point corresponding to the initial request information in the knowledge base by similarity calculation;

[0092] The reasoning module 430 is used to perform reasoning processing according to the pre-established knowledge graph when the corresponding knowledge point is not found in the knowledge base. The reasoning module 430 includes:

[0093] An instance query unit 431, configured to query a first instance in a pre-established knowledge graph;

[0094] An association path acquisition unit 432 is used to acquire an association path of the first instance in the knowledge graph until the association path includes a third instance;

[0095] The answer determination unit 433 is used to determine the answer according to the first instance, the second instance and the association path, where the association path includes M attribute information, where M is greater than or equal to 2.

[0096] In this embodiment, the output module can either return the text answer directly to the user, or convert the text answer into a voice format and then send it to the user, which does not limit the protection scope of the present invention.

[0097] Specifically, the knowledge graph can record the association relationship between multiple instances in the form of a table.

[0098] Specifically, the initial request information may include: what is the relationship between the first instance and the third instance or what is the relationship between the first instance and the third instance, etc.

[0099] Specifically, the knowledge point search module calculates the semantic similarity between the initial request information and the questions stored in the knowledge base, and compares the calculated maximum similarity with a preset first preset threshold. When the calculated similarity is greater than the first preset threshold, the knowledge point corresponding to the similarity is used as the knowledge point to be searched; otherwise, the corresponding knowledge point is not found in the knowledge base.

[0100] The technical solution of this embodiment receives the user's initial request information through a user request receiving module, and the initial request information includes a first instance and a third instance; the knowledge point search module searches for knowledge points corresponding to the initial request information in the knowledge base through similarity calculation; when the corresponding knowledge point is not found in the knowledge base, the reasoning module performs reasoning processing according to a pre-established knowledge graph; the reasoning module includes: an instance query unit, which is used to query the first instance in the pre-established knowledge graph; an associated path acquisition unit, which is used to obtain the associated path of the first instance in the knowledge graph until the associated path includes the third instance; an answer determination unit, which is used to determine the answer according to the first instance, the third instance and the associated path, so as to achieve finding the answer corresponding to the request information including two instances without changing the knowledge points stored in the knowledge base, and improve the accuracy of question and answer and save storage space without changing the content stored in the knowledge base.

[0101] Embodiment 3

[0102] Figure 3 A schematic diagram of the structure of a server provided in Embodiment 3 of the present invention is shown in FIG. Figure 3 As shown, the server includes a processor 510, a storage device 520, an input device 530, and an output device 540; the number of processors 510 in the server can be one or more. Figure 3 A processor 510 is taken as an example; the processor 510, the storage device 520, the input device 530 and the output device 540 in the server can be connected via a bus or other means. Figure 3 The example of connecting through bus is taken in the following.

[0103] The storage device 520, as a computer-readable storage medium, can be used to store software programs, computer executable programs and modules, such as program instructions / modules corresponding to the method for intelligent automatic question answering in Embodiment 1 of the present invention. The processor 510 executes various functional applications and data processing of the terminal device by running the software programs, instructions and modules stored in the storage device 520, that is, implements the above-mentioned method for intelligent automatic question answering.

[0104] The storage device 520 may mainly include a program storage area and a data storage area, wherein the program storage area may store an operating system and at least one application required for a function; the data storage area may store data created according to the use of the terminal, etc. In addition, the storage device 520 may include a high-speed random access memory, and may also include a non-volatile memory, such as at least one disk storage device, a flash memory device, or other non-volatile solid-state storage device. In some instances, the storage device 520 may further include a memory remotely arranged relative to the processor 510, and these remote memories may be connected to the terminal device via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0105] The input device 530 may be used to receive input digital or character information and generate key signal input related to user settings and function control of the terminal device. The output device 540 may include a display device such as a display screen.

[0106] Embodiment 4

[0107] Embodiment 4 of the present invention provides a storage medium containing computer executable instructions, and the computer executable instructions are used to execute a method of intelligent automatic question answering when executed by a computer processor. Please refer to embodiment 1 for details of this method, which will not be repeated here.

[0108] Through the above description of the implementation methods, the technicians in the relevant field can clearly understand that the present invention can be implemented by means of software and necessary general hardware, and of course it can also be implemented by hardware, but in many cases the former is a better implementation method. Based on such an understanding, the technical solution of the present invention is essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as a computer floppy disk, read-only memory (ROM), random access memory (RAM), flash memory (FLASH), hard disk or optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, server, or network device, etc.) to execute the methods described in each embodiment of the present invention.

[0109] It is worth noting that in the embodiment of the above-mentioned device, the various units and modules included are only divided according to functional logic, but are not limited to the above-mentioned division, as long as the corresponding functions can be achieved; in addition, the specific names of the functional units are only for the convenience of distinguishing each other, and are not used to limit the scope of protection of the present invention.

[0110] Note that the above are only preferred embodiments of the present invention and the technical principles used. Those skilled in the art will understand that the present invention is not limited to the specific embodiments described herein, and that various obvious changes, readjustments and substitutions can be made by those skilled in the art without departing from the scope of protection of the present invention. Therefore, although the present invention has been described in more detail through the above embodiments, the present invention is not limited to the above embodiments, and may include more other equivalent embodiments without departing from the concept of the present invention, and the scope of the present invention is determined by the scope of the appended claims.

Claims

1. A method for intelligent automatic question answering, characterized in that: include: Receiving initial request information from a user, wherein the initial request information includes a first instance and a third instance; Searching for a knowledge point corresponding to the initial request information in the knowledge base through similarity calculation; When the corresponding knowledge point is not found in the knowledge base, reasoning processing is performed according to the pre-established knowledge graph, and the reasoning processing includes: Querying a first instance in a pre-established knowledge graph; Obtaining an association path of the first instance in the knowledge graph until the association path includes a third instance; Determine an answer according to the first instance, the third instance and the association path, wherein the association path includes M attribute information, where M is greater than or equal to 2; Send the confirmed answer to the user; The knowledge graph records the association relationship between multiple instances in the form of a table; Searching for the knowledge point corresponding to the initial request information in the knowledge base through similarity calculation includes: calculating the semantic similarity between the initial request information and the question stored in the knowledge base, comparing the calculated maximum similarity with a preset first preset threshold, and when the calculated similarity is greater than the first preset threshold, using the knowledge point corresponding to the similarity as the knowledge point to be searched; otherwise, the corresponding knowledge point is not found in the knowledge base.

2. The method according to claim 1, characterized in that The initial request information includes: what is the relationship between the first instance and the third instance or what is the relationship between the first instance and the third instance.

3. An intelligent automatic question-answering system, characterized in that: include: A user request receiving module, used to receive initial request information of a user, wherein the initial request information includes a first instance and a third instance; A knowledge point search module, used to search the knowledge point corresponding to the initial request information in the knowledge base through similarity calculation; The reasoning module is used to perform reasoning processing according to the pre-established knowledge graph when the corresponding knowledge point is not found in the knowledge base. The reasoning module includes: An instance query unit, configured to query a first instance in a pre-established knowledge graph; an association path acquisition unit, configured to acquire an association path of the first instance in the knowledge graph until the association path includes a third instance; an answer determination unit, configured to determine an answer according to the first instance, the third instance, and the association path, wherein the association path includes M attribute information, where M is greater than or equal to 2; An output module, used to send the determined answer to the user; The knowledge graph records the association relationship between multiple instances in the form of a table; The knowledge point search module calculates the semantic similarity between the initial request information and the questions stored in the knowledge base, compares the calculated maximum similarity with a preset first preset threshold, and when the calculated similarity is greater than the first preset threshold, the knowledge point corresponding to the similarity is used as the knowledge point to be searched; otherwise, the corresponding knowledge point is not found in the knowledge base.

4. The system according to claim 3, characterized in that The initial request information includes: what is the relationship between the first instance and the third instance or what is the relationship between the first instance and the third instance.

5. A server, characterized in that: The server comprises: one or more processors; a storage device for storing one or more programs, When the one or more programs are executed by the one or more processors, the one or more processors implement the intelligent automatic question-answering method as described in any one of claims 1-2.

6. A computer storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the intelligent automatic question-answering method as described in any one of claims 1-2 is implemented.

Images