Method and Device for Processing Knowledge Data

Abstract

Embodiments of the present invention provide a method and device for processing knowledge data. The method includes: obtaining knowledge data to be detected; analyzing the knowledge data to be detected to obtain structured knowledge of the knowledge data to be detected; and detecting whether there is a conflict between the knowledge data to be detected and existing reference knowledge data. Wherein, the detection includes comparing the structured knowledge of the knowledge data to be detected with the structured knowledge obtained based on knowledge reasoning from the structured knowledge of two or more existing reference knowledge data to determine whether there is a conflict between the knowledge data to be detected and the existing reference knowledge data. The knowledge data processing method according to the embodiments of the present invention can improve the accuracy of conflict detection.

Description

Technical Field

[0001] The present invention generally relates to the technical field of data processing, and particularly to a method and device for processing knowledge data. Background Art

[0002] With the development of the Internet, people's channels for obtaining knowledge have become increasingly rich. In addition to the expert knowledge bases in various fields, various Internet knowledge bases have emerged, such as Wikipedia, Baidu Encyclopedia, etc. Expert knowledge bases mainly come from the personal experience of domain experts and face knowledge updates with the development of technology. Internet knowledge bases are established by Internet users and may contain incorrect knowledge. For knowledge from different data sources, there may be differences between knowledge data on the same knowledge topic, and there may even be conflicts or errors with each other.

[0003] Therefore, when constructing a knowledge base using knowledge data from multiple knowledge data sources, it is necessary to process the knowledge data therein, detect conflicts between the knowledge data, and exclude incorrect knowledge.

[0004] When detecting knowledge conflicts in the existing technology, it usually considers comparing new knowledge with the existing knowledge in the knowledge base one by one, but does not consider the conflicts between new knowledge and combinations of multiple pieces of knowledge in the knowledge base. Therefore, the existing patents cannot detect all conflicts between new knowledge and the knowledge in the knowledge base, resulting in a low accuracy rate of the knowledge data in the knowledge base.

[0005] Therefore, a mechanism for processing knowledge data with higher accuracy is needed. Summary of the Invention

[0006] To overcome at least some defects of the above-mentioned existing technology, embodiments of the present invention propose a method and device for processing knowledge data. When detecting knowledge conflicts, it not only considers whether there are conflicts between new knowledge and each piece of existing knowledge in the knowledge base, but also considers whether there are conflicts between new knowledge and combinations of multiple pieces of knowledge in the knowledge base. Therefore, the accuracy rate of conflict detection is improved. Correspondingly, the accuracy rate of the knowledge data in the built knowledge base can be improved.

[0007] According to a first aspect of the present invention, there is provided a method for processing knowledge data. The method includes: obtaining knowledge data to be detected; analyzing the knowledge data to be detected to obtain the structured knowledge of the knowledge data to be detected; and detecting whether there is a conflict between the knowledge data to be detected and the existing reference knowledge data. Wherein, the detection includes comparing the structured knowledge of the knowledge data to be detected with the structured knowledge obtained by knowledge reasoning based on the structured knowledge of two or more pieces of existing reference knowledge data to determine whether there is a conflict between the knowledge data to be detected and the existing reference knowledge data.

[0008] In some embodiments, the method further includes: before detecting whether there is a conflict between the knowledge data to be detected and the existing reference knowledge data, determining whether the structured knowledge of the knowledge data to be detected meets the preset attribute constraint conditions according to the preset attribute constraint detection rules.

[0009] In some embodiments, the knowledge data includes causal knowledge data, the knowledge reasoning includes causal knowledge reasoning, and the conflict includes causal relationship conflicts.

[0010] In some embodiments, the causal knowledge data includes at least one of the following:

[0011] A→B, indicating that A is the direct cause of B;

[0012] indicating that A is not the direct cause of B;

[0013] A-B, indicating that there is a direct causal relationship between A and B;

[0014] A⊥B, indicating that A and B do not affect each other;

[0015] indicating that A affects B;

[0016] indicating that A does not affect B;

[0017] A~B, indicating that A and B are related;

[0018] A≤B, indicating that the order of A is prior to B in the causal chain,

[0019] where A represents the subject in the causal knowledge data, B represents the object in the causal knowledge data, and the symbol between A and B represents the predicate in the causal knowledge data.

[0020] In some embodiments, the causal knowledge reasoning includes at least one of the following:

[0021] Based on the existing reference knowledge data A→B and B→C, reasoning to obtain A→B, B→C and

[0022] Based on the existing reference knowledge data A→B and A-B, reasoning to obtain A→B;

[0023] Based on the existing reference knowledge data A→B and B-A, reasoning to obtain A→B;

[0024] Based on the existing reference knowledge data A→B and reasoning to obtain A→B, and

[0025] Based on the existing reference knowledge data A→B and reasoning obtains A→B, and

[0026] Based on the existing reference knowledge data A→B and A~B, reasoning obtains A→B;

[0027] Based on the existing reference knowledge data A→B and B~A, reasoning obtains A→B;

[0028] Based on the existing reference knowledge data A→B and A≤B, reasoning obtains A→B;

[0029] Based on the existing reference knowledge data A→B and C≤A, reasoning obtains A→B, C≤A and C≤B;

[0030] Based on the existing reference knowledge data and A - B, reasoning obtains B→A;

[0031] Based on the existing reference knowledge data and B - A, reasoning obtains B→A;

[0032] Based on the existing reference knowledge data and A⊥B, reasoning obtains A⊥B;

[0033] Based on the existing reference knowledge data and B⊥A, reasoning obtains B⊥A;

[0034] Based on the existing reference knowledge data A - B and reasoning obtains and A→B;

[0035] Based on the existing reference knowledge data A - B and reasoning obtains and B→A;

[0036] Based on the existing reference knowledge data A - B and reasoning obtains B→A;

[0037] Based on the existing reference knowledge data A - B and reasoning obtains A→B;

[0038] Based on the existing reference knowledge data A - B and A≤B, reasoning obtains A→B;

[0039] Based on the existing reference knowledge data A - B and B≤A, reasoning obtains B→A;

[0040] Based on the existing reference knowledge data and reasoning obtains and

[0041] Based on the existing reference knowledge data and A to B, infer to obtain

[0042] Based on the existing reference knowledge data and B to A, infer to obtain

[0043] Based on the existing reference knowledge data and A ≤ B, infer to obtain

[0044] Based on the existing reference knowledge data and C ≤ A, infer to obtain C ≤ A and C ≤ B; and

[0045] Based on the existing reference knowledge data A ≤ B and B ≤ C, infer to obtain A ≤ B, B ≤ C and A ≤ C.

[0046] In some embodiments, causal relationship conflicts include at least one of the following:

[0047] A → B and B → A, A ⊥ B, B ⊥ A, and any one of the conflicts in B ≤ A;

[0048] conflicts with A → B;

[0049] A - B conflicts with any one of A ⊥ B and B ⊥ A;

[0050] A ⊥ B conflicts with any one of A → B, B → A, A - B, B - A, A to B, and B to A;

[0051] conflicts with B → A, conflicts with any one of B ≤ A, A ⊥ B and B ⊥ A;

[0052] conflicts with A → B and any one of;

[0053] A to B conflicts with any one of A ⊥ B and B ⊥ A;

[0054] A ≤ B conflicts with B → A and any one of;

[0055] According to a second aspect of the present invention, there is provided a processing device for knowledge data. The processing device includes: one or more processors; and 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 are configured to execute the above method.

[0056] According to a third aspect of the present invention, there is provided a computer-readable storage medium having computer instructions stored thereon, and when the instructions are executed by a processor, the steps of the above-described method are implemented. BRIEF DESCRIPTION OF THE DRAWINGS

[0057] The preferred embodiments of the present disclosure will be described below with reference to the accompanying drawings, which will make the above and other objects, features and advantages of the present invention clearer. Among them:

[0058] Figure 1 FIG. shows a schematic flow chart of a method for processing knowledge data according to an embodiment of the present invention;

[0059] Figure 2 FIG. shows a schematic flow chart of a method for processing knowledge data according to another embodiment of the present invention;

[0060] Figure 3 FIG. shows a schematic block diagram of a device for processing knowledge data according to an embodiment of the present invention.

[0061] In all the drawings of the present disclosure, the same or similar reference numerals denote the same or similar elements. DETAILED DESCRIPTION

[0062] The principles and spirit of the present disclosure will be described below with reference to several exemplary embodiments in conjunction with the accompanying drawings. It should be understood that these embodiments are given only to enable those skilled in the art to better understand and thus implement the present disclosure, and do not limit the scope of the present disclosure in any way. In addition, for the sake of simplicity, the detailed description of well-known technologies not directly related to the present invention is omitted to prevent confusion in the understanding of the present invention.

[0063] The terms used herein are only for describing exemplary embodiments and are not intended to limit the exemplary embodiments. As used herein, unless clearly indicated in the context, the singular form does not exclude the possibility of also including the plural form. It should also be understood that when used in this specification, "and / or" includes any and all combinations of one or more of the related listed items. The terms "comprising" and / or "having" specify the presence of the recited features, numbers, steps, operations, components, elements or combinations thereof, and do not exclude the presence or addition of one or more other features, numbers, steps, operations, components, elements or combinations thereof.

[0064] Unless otherwise clearly defined, all terms used herein have the same meaning as commonly understood by those of ordinary skill in the art to which the exemplary embodiments belong. It should also be understood that, unless otherwise clearly defined herein, terms should be interpreted as having a meaning consistent with the meaning understood by those of ordinary skill in the art in the specification at the time of the invention.

[0065] Embodiments of the present invention provide a method and device for processing knowledge data. When detecting knowledge conflicts, it not only considers whether there are conflicts between new knowledge and each piece of existing knowledge in the knowledge base, but also considers whether there are conflicts between new knowledge and combinations of multiple pieces of knowledge in the knowledge base. Therefore, the accuracy of detecting data conflicts is improved. Correspondingly, the accuracy rate of the knowledge data in the knowledge base constructed using the knowledge data processing method of the embodiments of the present invention is improved.

[0066] For ease of understanding, several terms used in the present invention are briefly described below. In the present disclosure, a knowledge base refers to a collection of interrelated knowledge stored, organized, managed, and used in a computer memory by adopting a certain (or several) knowledge representation method for the need of solving problems in a certain (or certain) field. Conflict detection refers to detecting whether there are contradictions between different pieces of knowledge.

[0067] Figure 1 A schematic flowchart of a method 100 for processing knowledge data according to an embodiment of the present invention is shown.

[0068] As mentioned above, when constructing a knowledge base using the knowledge data in multiple knowledge data sources, it is necessary to process the knowledge data therein, detect the conflicts between the knowledge data, and exclude the incorrect knowledge. In this case, the method 100 according to an embodiment of the present invention can be used.

[0069] As shown in the figure, in step S110, the knowledge data to be detected is obtained. The knowledge data to be detected can be obtained from various knowledge data sources. The knowledge data sources can be, for example, various knowledge bases, such as expert knowledge bases, Internet knowledge bases, and so on.

[0070] The knowledge data can be one or more statements in the knowledge base. Examples of the knowledge data are such as "Nanjing is located in the eastern part of China, downstream of the Yangtze River, and is the capital city of Jiangsu Province", and also such as "Temperature directly affects the sales volume of beer".

[0071] In step S120, the knowledge data to be detected is analyzed to obtain the structured knowledge of the knowledge data to be detected.

[0072] In some instances, the data of the subject, predicate, and object can be extracted from the knowledge data to be detected, and the structured knowledge of the knowledge data to be detected can be obtained. For example, the subject can be the subject in the knowledge data, the predicate can be the predicate in the knowledge data, and the object can be the object in the knowledge data. The structured knowledge of the knowledge data can be the knowledge data composed of the subject, predicate, and object. For example, a piece of knowledge such as "Nanjing is located in the eastern part of China, downstream of the Yangtze River, and is the capital city of Jiangsu Province" can be structured as (Nanjing, is, the capital city of Jiangsu Province).

[0073] In some instances, nouns and the causal relationships between the nouns can be extracted from the knowledge data to be detected, and the structured knowledge of the knowledge data to be detected (such knowledge is also called causal knowledge data) can be obtained. For example, a piece of knowledge such as "Temperature directly affects the sales volume of beer" can be structured as (Temperature, directly affects, the sales volume of beer).

[0074] In step S130, it is detected whether there is a conflict between the knowledge data to be detected and the existing reference knowledge data.

[0075] The existing reference knowledge data can be the knowledge data that has been confirmed as accurate. The existing reference knowledge data can be selected from the knowledge data existing in the built knowledge base. Optionally, the existing reference knowledge data can be selected from the knowledge data of the existing knowledge base without conflict (such as an expert knowledge base) serving as a reference source.

[0076] In particular, the detection in step S130 can include comparing the structured knowledge of the knowledge data to be detected with the structured knowledge obtained based on knowledge reasoning from the structured knowledge of two or more existing reference knowledge data to determine whether there is a conflict between the knowledge data to be detected and the existing reference knowledge data. In other words, in the embodiments of the present invention, when detecting conflicts, the reference objects compared with the knowledge data to be detected include not only each piece of existing reference knowledge data itself, but also the knowledge data obtained by reasoning from two or more existing reference knowledge data based on the knowledge reasoning algorithm. It should be understood that in this article, when referring to the comparison between knowledge data, it refers to the comparison between the structured knowledge of the respective knowledge data, rather than the comparison between the original texts of the knowledge data. This is only for the sake of simplicity of description and will not be elaborated further below.

[0077] Knowledge data includes various types. For specific types of knowledge data, there can be corresponding conflict detection rules.

[0078] In an instance of detecting conflicts between structured knowledge composed of a subject, a predicate, and an object, when the content information of any two of the subject, predicate, and object of two pieces of knowledge is the same, and the content information of the remaining one is different, it can be considered that there is a conflict between these two pieces of knowledge.

[0079] Optionally, in an instance of detecting a conflict between causal relationship data, when there is a conflict between the causal relationships contained in two pieces of knowledge, it can be considered that there is a conflict between these two pieces of knowledge.

[0080] The conflict detection in step S130 will be mainly described below by taking causal knowledge data as an example. For the sake of easy understanding, in such an embodiment, both the knowledge data to be detected and the existing reference knowledge data are causal knowledge data. Unless otherwise explicitly stated below, knowledge or knowledge data generally refers to causal knowledge data. The existing reference knowledge data can be obtained by reasoning a reference knowledge data set based on a knowledge reasoning algorithm such as causal knowledge reasoning. When there is a causal relationship conflict between two pieces of causal knowledge data, it can be determined that there is a conflict between them.

[0081] For causal knowledge data, according to the causal relationship predicate, it can be classified into 8 types, as described in Table 1 below.

[0082]

[0083]

[0084] Table 1

[0085] For the selected multiple pieces of reference knowledge, causal knowledge reasoning can be performed according to the reasoning rules in Table 2 below to obtain a reference knowledge data set.

[0086]

[0087]

[0088]

[0089] Table 2

[0090] Table 2 shows the causal knowledge data set (as shown in the column of reasoning result) obtained by causal knowledge reasoning from two existing pieces of causal knowledge data (as shown in the column of existing causal relationship).

[0091] It is easy to understand that based on the rules in Table 2, through the principle of pairwise combination, a causal knowledge data set obtained by reasoning from three or more existing pieces of causal knowledge data can be deduced. For example, according to three existing pieces of causal knowledge data: A→B, B→C, C→D, it can be deduced that A→B, B→C, C→D.

[0092] Each type of causal knowledge has a corresponding conflict set. Table 3 below shows each type of causal knowledge and its conflict set.

[0093]

[0094]

[0095] Table 3

[0096] In step S130, to detect whether there is a conflict between the knowledge data to be detected and the existing reference knowledge data, it can be determined by detecting whether the knowledge data to be detected is in the conflict set of the existing reference knowledge data and its inference results. If the knowledge to be detected is in the conflict set of any existing reference knowledge data or its inference results, it can be determined that there is a conflict between the knowledge data to be detected and the existing reference knowledge data.

[0097] It should be understood that the method 100 according to the embodiments of the present invention is not limited to the steps and sequences shown above.

[0098] Optionally, before step S130, a pre-detection step can be performed first. In this pre-detection step, according to the preset attribute constraint detection rules, it can be determined whether the structured knowledge of the knowledge data to be detected satisfies the preset attribute constraint conditions. For example, when the attribute information of the subject or object is age, it needs to satisfy the constraint conditions (age > 0 and age < 200).

[0099] Optionally, in step S130, when the existing reference knowledge data is selected from the knowledge data already existing in the knowledge base being built, the knowledge base can be traversed to detect whether there is a conflict between the knowledge data to be detected and the existing knowledge data.

[0100] The method 100 according to the embodiments of the present invention not only considers whether there is a conflict between the new knowledge and each piece of existing knowledge in the knowledge base, but also considers whether there is a conflict between the new knowledge and the combination of multiple pieces of knowledge in the knowledge base. Therefore, more conflicting knowledge data can be discovered, and the conflict detection accuracy is improved.

[0101] Next, refer to Figure 2 to show a preferred embodiment of the present invention. Figure 2 FIG. shows a schematic flowchart of a method 200 for processing knowledge data according to another embodiment of the present invention. The method 200 can be used to build a knowledge base with high accuracy.

[0102] As shown in the figure, in step S210, the knowledge data to be detected (hereinafter also referred to as new knowledge) can be obtained.

[0103] Specifically, the knowledge data to be detected can be obtained from a certain source, such as an expert knowledge base, an Internet knowledge base, etc.

[0104] In step S220, the knowledge data to be detected can be processed into structured knowledge.

[0105] Specifically, the data of the subject, predicate, and object are extracted from the knowledge data to be detected, and the corresponding structured knowledge data is obtained. For example, a piece of knowledge such as "Temperature directly affects the sales volume of beer" can be structured as (temperature, directly affects, sales volume of beer), and marked as m→n, where m represents temperature, n represents the sales volume of beer, and → represents that the former directly affects the latter.

[0106] In the optional step S230, it can be determined whether the attribute constraint conditions are met.

[0107] Specifically, the attribute information of the subject and object in the knowledge data to be detected can be analyzed according to the preset attribute constraint detection rules to determine whether the knowledge data to be detected meets the preset attribute constraint conditions. For example, when the attribute information is age, the constraint conditions (age > 0 and age < 200) need to be met.

[0108] If it is determined in step S230 that the attribute constraint conditions are not met, the method proceeds to step S240. In step S240, an attribute error prompt message is output.

[0109] If it is determined in step S230 that the attribute constraint conditions are met, the method proceeds to step S250. In step S250, several existing causal knowledge data are selected from the knowledge base under construction as reference knowledge. Since there may be conflicts between the knowledge data to be detected and the combinations of multiple existing knowledge in the knowledge base, it is necessary to select two or more existing knowledge from the database at one time in order to determine whether the knowledge data to be detected conflicts with these combinations of knowledge in the following steps.

[0110] In step S260, for the selected multiple existing knowledge, according to the causal knowledge reasoning, the corresponding knowledge set is obtained. For example, when two pieces of knowledge are selected at one time, the causal knowledge reasoning can be directly carried out in the way of Table 2. When three or more pieces of knowledge are selected at one time, as described above, according to the principle of Table 2, for example, by pairwise combination, the corresponding knowledge set can be obtained through reasoning.

[0111] As an example, if the existing knowledge in the knowledge base is and p≤m, then according to inference rule 24, p≤n can be inferred, thus obtaining the knowledge set p≤m and p≤n.

[0112] In step S270, according to the conflict detection algorithm, it is detected whether there is a conflict between the knowledge data to be detected and the existing knowledge data.

[0113] As shown in Table 3, each type of causal knowledge data has a corresponding conflict set. When performing conflict detection, it is necessary to detect one by one whether the new knowledge is in the conflict sets of several pieces of knowledge selected in step S250 and their reasoning results (obtained in step S260).

[0114] For example, if the new knowledge is Two pieces of existing knowledge selected from the knowledge base in step S250 are and p ≤ m. In step S260, from these two pieces of knowledge, p ≤ n is deduced to obtain the knowledge set p ≤ m and p ≤ n. Then, when performing conflict detection in step S270, it is necessary to detect in sequence whether it is in the conflict set of the three pieces of knowledge p ≤ m, p ≤ n. In this example, since is in the conflict set of p ≤ n, there is a conflict.

[0115] If a conflict is detected in step S270, the method proceeds to step S280. In step S280, a conflict prompt message is output.

[0116] If no conflict is detected in step S280, the method proceeds to step S290. In step S290, it is judged whether the entire knowledge base has been traversed. If not, the method returns to step S250, and several other pieces of existing causal knowledge data are selected from the knowledge base under construction as reference knowledge. Then, continue with the aforementioned steps S260 - 290 to detect whether the knowledge data to be detected conflicts with the several pieces of existing causal knowledge data selected this time.

[0117] If it is judged in step S290 that the entire knowledge base has been traversed, the method proceeds to step S295, and the knowledge data that has been detected is added to the database.

[0118] When the method 200 according to the embodiment of the present invention detects knowledge conflicts, based on the knowledge reasoning algorithm, it considers whether there are conflicts between the new knowledge to be detected to be added to the knowledge base and various combinations of all the knowledge in the knowledge base, and traverses various combinations of the knowledge base. Based on the knowledge processing method of the present disclosure, all conflicts between the new knowledge and the knowledge in the knowledge base can be detected, the construction of the knowledge base can be successfully realized, and the accuracy rate of the knowledge base can be improved.

[0119] Figure 3 A schematic block diagram of a processing device 300 for knowledge data according to an embodiment of the present invention is schematically shown.

[0120] As Figure 3As shown, the processing device includes a processing unit or processor 136. The processor 136 can be a single unit or a combination of multiple units, and is used to perform various operations, including the method of processing knowledge data according to the embodiments of the present invention, such as method 100 or 200. The processing device 300 may further include: an input unit 132 for receiving signals from other devices or components; and an output unit 134 for providing signals to other devices or components. The input unit and the output unit can be arranged as a whole.

[0121] In addition, as shown in the figure, the processing device 300 further includes one or more storage devices 138, and a computer program 139 is stored in the storage device 138.

[0122] The computer program 139 may include code / computer-executable instructions, which, when executed by the processor 136, cause the processor 136 to perform, for example, the operation flow of the method described above in conjunction with Figures 1 to 2 and any variations thereof.

[0123] The computer program 139 can be configured to have computer program code including, for example, computer program modules. For example, in an exemplary embodiment, the code in the computer program 139 may include one or more program modules, such as including 139A, module 139B,.... It should be noted that the way of dividing the modules and the number of modules are not fixed, and those skilled in the art can use appropriate program modules or combinations of program modules according to the actual situation. When these combinations of program modules are executed by the processor 136, the processor 136 can perform, for example, the method flow described above in conjunction with Figure 1 and Figure 2 and any variations thereof.

[0124] The present invention has been described above in conjunction with preferred embodiments. The method for processing knowledge data according to the embodiments of the present invention can detect all conflicts between new knowledge and the knowledge in the knowledge base, successfully realize the construction of the knowledge base, and improve the accuracy of the knowledge base.

[0125] It can be understood that the devices and methods shown above are only exemplary. The devices of the present invention may include more or fewer components than those shown. The methods of the present invention are not limited to the steps and sequences shown above. Those skilled in the art can make many changes and modifications according to the teachings of the illustrated embodiments.

[0126] The above-mentioned methods, devices, units, and / or modules according to the embodiments of the present application can be implemented by a computing-capable electronic device executing software containing computer instructions. The system may include a storage device to implement the various storages described above. The computing-capable electronic device may include a general-purpose processor, a digital signal processor, a dedicated processor, a reconfigurable processor, or other devices capable of executing computer instructions, but is not limited thereto. Executing such instructions configures the electronic device to perform the above-mentioned operations according to the present application. The above-mentioned devices and / or modules may be implemented in one electronic device or in different electronic devices. Such software may be stored in a computer-readable storage medium. The computer-readable storage medium stores one or more programs (software modules), and the one or more programs include instructions that, when executed by one or more processors in the electronic device, cause the electronic device to execute the method of the present application.

[0127] These software can be stored in the form of volatile memory or non-volatile storage devices (such as storage devices like ROM), whether erasable or rewritable, or stored in the form of memory (such as RAM, memory chips, devices, or integrated circuits), or stored on an optically readable medium or a magnetically readable medium (such as, CD, DVD, disk, or magnetic tape, etc.). It should be realized that the storage device and the storage medium are embodiments of machine-readable storage devices suitable for storing one or more programs, and the one program or more programs include instructions that, when executed, implement the embodiments of the present application. The embodiments provide a program and a machine-readable storage device storing such a program, and the program includes code for implementing the device or method described in any one of the claims of the present application. In addition, these programs can be transmitted by electrical means via any medium (such as a communication signal carried via a wired connection or a wireless connection), and multiple embodiments appropriately include these programs.

[0128] The methods, devices, units, and / or modules according to the embodiments of the present application can also be implemented using hardware or firmware such as a field-programmable gate array (FPGA), a programmable logic array (PLA), a system-on-chip, a system-on-substrate, a system-on-package, an application-specific integrated circuit (ASIC), or any other reasonable way suitable for integrating or packaging circuits, or implemented in an appropriate combination of the three implementation manners of software, hardware, and firmware. The system may include a storage device to implement the storage described above. When implemented in these ways, the software, hardware, and / or firmware used are programmed or designed to execute the corresponding above-mentioned methods, steps, and / or functions according to the present application. Those skilled in the art can appropriately implement one or more of these systems and modules, or a part or multiple parts thereof, using different above-mentioned implementation manners according to actual needs. These implementation manners all fall within the protection scope of the present application.

[0129] Although the present application has been shown and described with reference to specific exemplary embodiments thereof, those skilled in the art should understand that various changes in form and detail may be made therein without departing from the spirit and scope of the present application as defined by the appended claims and their equivalents. Therefore, the scope of the present application should not be limited to the above-described embodiments, but should be determined not only by the appended claims but also by the equivalents of the appended claims.

Claims

1. A method for processing knowledge data executed by an electronic device, including: obtaining knowledge data to be detected from a knowledge data source, analyzing the knowledge data to be detected to obtain structured knowledge of the knowledge data to be detected, detecting whether there is a conflict between the knowledge data to be detected and existing reference knowledge data, excluding incorrect knowledge according to the detected conflict, wherein the detection includes comparing the structured knowledge of the knowledge data to be detected with the structured knowledge obtained based on knowledge reasoning from the structured knowledge of two or more existing reference knowledge data to determine whether there is a conflict between the knowledge data to be detected and the existing reference knowledge data, wherein the knowledge data includes causal knowledge data, the knowledge reasoning includes causal knowledge reasoning, and the conflict includes causal relationship conflict.

2. The method according to claim 1, further including: determining whether the structured knowledge of the knowledge data to be detected meets a preset attribute constraint condition according to a preset attribute constraint detection rule.

3. The method according to claim 1 or 2, wherein the causal knowledge data includes at least one of the following: A→B, indicating that A is the direct cause of B; Indicates that A is not the direct cause of B; A-B, indicating that there is a direct causal relationship between A and B; A⊥B, indicating that A and B do not affect each other; It indicates that A affects B; It means that A will not affect B; A~B, indicating that A and B are related; A≤B, indicating that the order of A is prior to B in the causal chain, where A represents the subject in the causal knowledge data, B represents the object in the causal knowledge data, and the symbol between A and B represents the predicate in the causal knowledge data.

4. The method according to claim 3, wherein, the causal knowledge reasoning includes at least one of the following: Based on the existing reference knowledge data A→B and B→C, it is deduced that A→B, B→C and reasoning A→B based on existing reference knowledge data A→B and A-B; reasoning A→B based on existing reference knowledge data A→B and B-A; According to the existing reference knowledge data A→B and reason to obtain A→B, and Based on the existing reference knowledge data A→B and infer A→B, and reasoning A→B based on existing reference knowledge data A→B and A~B; reasoning A→B based on existing reference knowledge data A→B and B~A; reasoning A→B based on existing reference knowledge data A→B and A≤B; reasoning A→B, C≤A and C≤B based on existing reference knowledge data A→B and C≤A; Based on the existing reference knowledge data From A - B, it is deduced that B → A; Based on existing reference knowledge data From B - A, it is inferred that B → A; Based on the existing reference knowledge data Given A⊥B, it is deduced that A⊥B; Based on existing reference knowledge data and B⊥A, it is deduced that B⊥A; Based on the existing reference knowledge data A - B and infer to obtain and A → B; Based on the existing reference knowledge data A - B and infer to obtain and B → A; Based on the existing reference knowledge data A - B and infer B → A; Based on the existing reference knowledge data A - B and infer A → B; reasoning A→B based on existing reference knowledge data A-B and A≤B; reasoning B→A based on existing reference knowledge data A-B and B≤A; Based on existing reference knowledge data and reasoned to obtain and Based on existing reference knowledge data and A to B, infer and obtain Based on existing reference knowledge data and B~A, it is deduced that Based on existing reference knowledge data and A ≤ B, it is deduced that Based on existing reference knowledge data and C ≤ A, it is deduced that C ≤ A and C ≤ B; and reasoning A≤B, B≤C and A≤C based on existing reference knowledge data A≤B and B≤C.

5. The method according to claim 3, wherein, the causal relationship conflict includes at least one of the following: A→B and B→A, A⊥B, B⊥A, conflicts with any one of B≤A; Conflicts with A→B; A-B conflicts with any one of A⊥B and B⊥A; A⊥B conflicts with any one of A→B, B→A, A - B, B - A, A~B, and B~A; conflicts with any one of B→A, B≤A, A⊥B, and B⊥A; conflict with any one of A→B and in; A~B conflicts with any one of A⊥B and B⊥A; A ≤ B conflicts with any one of B → A and in it.

6. A knowledge data processing device, including: one or more processors; and a storage device for storing one or more programs, wherein, when the one or more programs are executed by the one or more processors, the one or more processors are configured to: obtain knowledge data to be detected from a knowledge data source, analyze the knowledge data to be detected to obtain structured knowledge of the knowledge data to be detected, Detect whether there is a conflict between the knowledge data to be detected and the existing reference knowledge data. Exclude the incorrect knowledge according to the detected conflict. Among them, the detection includes comparing the structured knowledge of the knowledge data to be detected with the structured knowledge obtained based on knowledge reasoning from the structured knowledge of two or more existing reference knowledge data to determine whether there is a conflict between the knowledge data to be detected and the existing reference knowledge data. Among them, the knowledge data includes causal knowledge data, the knowledge reasoning includes causal knowledge reasoning, and the conflict includes causal relationship conflicts.

7. The processing device according to claim 6, wherein the one or more processors are further configured to: Determine whether the structured knowledge of the knowledge data to be detected satisfies the preset attribute constraint conditions according to the preset attribute constraint detection rules.

8. The processing device according to claim 6 or 7, wherein the causal knowledge data includes at least one of the following: A→B, indicating that A is the direct cause of B; Indicates that A is not the direct cause of B; A-B, indicating that there is a direct causal relationship between A and B; A⊥B, indicating that A and B do not affect each other; It means that A affects B; It means that A will not affect B; A~B, indicating that A and B are related; A≤B, indicating that the order of A is prior to B in the causal chain, where A represents the subject in the causal knowledge data, B represents the object in the causal knowledge data, and the symbol between A and B represents the predicate in the causal knowledge data.

9. The processing device according to claim 8, wherein, the causal knowledge reasoning includes at least one of the following: Based on the existing reference knowledge data A→B and B→C, it is deduced that A→B, B→C and Based on the existing reference knowledge data A→B and A-B, infer A→B; Based on the existing reference knowledge data A→B and B-A, infer A→B; Based on the existing reference knowledge data A→B and infer A→B, and According to the existing reference knowledge data A→B and infer A→B, and Based on the existing reference knowledge data A→B and A~B, infer A→B; Based on the existing reference knowledge data A→B and B~A, infer A→B; Based on the existing reference knowledge data A→B and A≤B, infer A→B; Based on the existing reference knowledge data A→B and C≤A, infer A→B, C≤A and C≤B; Based on the existing reference knowledge data From A - B, it is inferred that B → A; Based on the existing reference knowledge data From B - A, it is deduced that B → A; Based on existing reference knowledge data Given A⊥B, it can be deduced that A⊥B; Based on the existing reference knowledge data and B⊥A, it is deduced that B⊥A; Based on the existing reference knowledge data A - B and reasoned to obtain and A → B; Based on the existing reference knowledge data A - B and infer to obtain and B → A; Based on the existing reference knowledge data A - B and infer B → A; Based on the existing reference knowledge data A - B and reason to obtain A → B; Based on the existing reference knowledge data A-B and A≤B, infer A→B; Based on the existing reference knowledge data A-B and B≤A, infer B→A; Based on existing reference knowledge data and inferred to obtain and Based on existing reference knowledge data and A to B, it is deduced that Based on the existing reference knowledge data and B~A, it is deduced that Based on existing reference knowledge data and A ≤ B, it is deduced that Based on the existing reference knowledge data and C ≤ A, it is deduced that C ≤ A and C ≤ B; and Based on the existing reference knowledge data A≤B and B≤C, infer A≤B, B≤C and A≤C.

10. The processing device according to claim 8, wherein, the causal relationship conflict includes at least one of the following: A→B and B→A, A⊥B, B⊥A, conflicts with any one of B≤A; Conflicts with A→B; A-B conflicts with any one of A⊥B and B⊥A; A⊥B conflicts with any one of A→B, B→A, A-B, B-A, A~B, and B~A; conflicts with any one of B→A, B≤A, A⊥B, and B⊥A; conflicts with any one of A→B and ; A~B conflicts with any one of A⊥B and B⊥A; A ≤ B conflicts with any one of B → A and in it.

11. A computer-readable storage medium, on which computer instructions are stored, and when the instructions are executed by a processor, the steps of the method according to any one of claims 1 to 5 are implemented.

Images