A method, device, and medium for constructing a large geological model

Abstract

The application provides a geological large model construction method and device, equipment and medium, relates to the geological technical field, and includes the following steps: obtaining geological data to be processed; performing entity identification on the geological data to obtain a plurality of geological entities and first associated data; for any geological entity, calculating the similarity of each geological term in the geological entity and a configured geological term dictionary; taking the geological term with the highest similarity to the geological entity as the target geological term of the geological entity; inputting the geological entity, the first associated data, and the similarity of the geological entity and the target geological term into a pre-trained relation extraction model to obtain second associated data; obtaining expert opinions of a geological expert; and constructing a geological large model based on the geological term dictionary, the geological entities, the corresponding first associated data and second associated data, the similarity to the corresponding target geological term, and the expert opinions. The application can construct a comprehensive and accurate geological large model.

Description

Technical Field

[0001] This application relates to the field of geological technology, and more specifically, to a method, apparatus, equipment, and medium for constructing a large geological model. Background Technology

[0002] With the advancement of the Digital Earth strategy, the geological field is undergoing a paradigm shift from experience-driven to data-intelligent-driven. The construction of large-scale geological models is a crucial aspect of geological research. Current technologies lack specialized large-scale geological models, and the content output by general-purpose models often contains misinformation. Geological data is mostly in unstructured forms such as text reports, charts, and logs, with entities, relationships, and attributes scattered and lacking systematic connections. Existing general-purpose models struggle to integrate this fragmented information into structured knowledge, resulting in low knowledge reuse rates. Simultaneously, new geological terms and discoveries are constantly emerging, requiring traditional general-purpose models to be retrained through large-scale manual annotation to adapt to new knowledge, which is costly and time-consuming. Summary of the Invention

[0003] The purpose of this application is to provide a method, apparatus, equipment and medium for constructing a large geological model, so as to solve the above-mentioned problems existing in the prior art, and to construct a comprehensive and accurate professional large geological model.

[0004] Firstly, a method for constructing a large geological model is provided, which may include: Obtain the geological data to be processed; Entity identification is performed on the geological data to obtain multiple geological entities and first association data of each geological entity; For any given geological entity, calculate the similarity between the geological entity and the geological terms in the configured geological terminology dictionary. The geological term with the highest similarity to the geological entity is taken as the target geological term for the geological entity. The geological entity, the first associated data of the geological entity, and the similarity between the geological entity and the target geological term are input into a pre-trained relation extraction model to obtain the second associated data corresponding to the geological entity. Obtain expert opinions from geological experts on the second correlation data of various geological entities; Based on the geological terminology dictionary, various geological entities, and corresponding first and second association data, similarity with corresponding target geological terms, and expert opinions, a large geological model is constructed.

[0005] In an optional implementation, the method for configuring the geological terminology dictionary includes: To obtain the associated data of different geological terms in the field of geology, as well as the metadata of academic materials containing the corresponding geological terms; For any given geological term, assign a unique term ID to that geological term; Calculate the overall authority score of the geological term based on the metadata of academic materials containing the geological term; If the overall authority score of the geological term is greater than the configured authority threshold, then the geological term will be used as the target geological term. A geological terminology dictionary is constructed based on the term IDs and associated data of different target geological terms.

[0006] In an optional implementation, the first associated data includes: the dependency syntax tree of the geological entity, its location, and a text block containing the corresponding geological entity; Entity identification was performed on the geological data to obtain multiple geological entities, including: The geological data to be processed is segmented, part-of-speech tagging is performed, and dependency parsing is performed to obtain multiple words, the part-of-speech tags of the corresponding words, and dependency parsing trees. Extract text blocks containing each word from the geological data; The text blocks containing each word, the part-of-speech tag and dependency syntax tree of each word are input into the pre-trained geological entity recognition model to obtain multiple geological entities and their corresponding dependency syntax trees, locations and text blocks containing the corresponding geological entities.

[0007] In an optional implementation, the relation extraction model includes: The text encoding layer is used to encode text blocks containing input geological entities and the locations of the geological entities to obtain context feature vectors and global semantic vectors of text blocks. The syntactic coding layer is used to encode the dependency paths of the input geological entities and their dependency syntax trees to obtain syntactic feature vectors. The rule encoding layer is used to encode the geological rules and geological terminology dictionary in the configured geological rule base to obtain rule feature vectors; The feature fusion and rule adjustment attention layer is used to concatenate the geological entity and its corresponding context feature vector, text block global semantic vector, syntactic feature vector and rule feature vector based on the similarity between the geological entity and the target geological term, and then perform linear transformation and attention weighting to obtain a weighted feature vector. The relationship classification layer is used to determine the second association data and corresponding confidence level of the geological entity based on the weighted feature vector; the second association data includes: the relationship type between the geological entity and other geological entities besides the geological entity, the relationship type between the geological entity and the target geological term, and the attribute information of the geological entity.

[0008] In an optional implementation, a large geological model is constructed based on the geological terminology dictionary, various geological entities, and corresponding first and second association data, similarity with corresponding target geological terms, and expert opinions, including: Based on the expert opinions and the configured geological rule base, the second correlation data of each geological entity is corrected to obtain the third correlation data of each geological entity. Based on the geological rule base, the geological terminology dictionary, various geological entities, and the corresponding first and third related data and similarity with the corresponding target geological terms, an initial geological knowledge graph is constructed. Using the Datalog inference engine, the geological rule base, the initial geological knowledge graph, and the geological terminology dictionary are updated to obtain the geological knowledge graph and the updated geological rule base and geological terminology dictionary; The geological knowledge graph is used to train the pre-constructed initial geological model to obtain the large geological model.

[0009] In an optional implementation, a geological knowledge graph is constructed based on the geological rule base, the geological terminology dictionary, various geological entities, and corresponding first and third association data, as well as the similarity with the corresponding target geological terms. This includes: Using geological terms and entities from the geological terminology dictionary as nodes, edges are established between nodes based on the geological rule base, the third-party associated data of the corresponding geological entities, and the similarity with the corresponding target geological terms. Attribute information is added to the nodes. Based on the source of the text blocks containing the corresponding geological entities, the attribute information of the nodes is associated with the attribute source to obtain a geological knowledge graph.

[0010] In an optional implementation, updating the geological rule base, the initial geological knowledge graph, and the geological terminology dictionary includes: From the geological rule base, match the first geological rule corresponding to the target geological term and the second geological rule corresponding to the relationship type; The Datalog inference engine is used to perform rule inference on the first geological rule and the second geological rule to obtain multiple implicit geological rules; Calculate the confidence level of each implicit geological rule, and perform consistency verification between each implicit geological rule and the initial geological knowledge map; Obtain the modification opinions of geological experts on each implicit geological rule based on the confidence level and consistency verification results; Based on the acquired implicit geological rules and corresponding correction suggestions, the geological rule base, the initial geological knowledge graph, and the geological terminology dictionary are updated to obtain the updated geological rule base, geological knowledge graph, and geological terminology dictionary.

[0011] Secondly, a device for constructing a large geological model is provided, which may include: The first acquisition unit is used to acquire geological data to be processed. The identification unit is used to perform entity identification on the geological data to obtain multiple geological entities and first association data of each geological entity; The calculation unit is used to calculate the similarity between any geological entity and various geological terms in a configured geological terminology dictionary for any given geological entity; and to select the geological term with the highest similarity to the geological entity as the target geological term for the geological entity. The input unit is used to input the geological entity, the first associated data of the geological entity, and the similarity between the geological entity and the target geological term into a pre-trained relation extraction model to obtain the second associated data corresponding to the geological entity. The second acquisition unit is used to acquire expert opinions from geological experts on the second correlation data of various geological entities; The construction unit is used to construct a large geological model based on the geological terminology dictionary, various geological entities, and corresponding first and second related data, similarity with corresponding target geological terms, and expert opinions.

[0012] Thirdly, an electronic device is provided, which includes a processor, a communication interface, a memory, and a communication bus, wherein the processor, the communication interface, and the memory communicate with each other through the communication bus; Memory, used to store computer programs; When a processor executes a program stored in memory, it implements any of the steps described in the first aspect above.

[0013] Fourthly, a computer-readable storage medium is provided, wherein a computer program is stored therein, and when executed by a processor, the computer program implements the steps of any of the methods described in the first aspect above.

[0014] This application extracts entities from unstructured geological data and establishes relationships between these entities and their first and second related data. It then integrates this information with a geological terminology dictionary to achieve standardized integration, transforming scattered geological knowledge into a relational knowledge graph. Furthermore, by calculating the similarity between geological entities and terms in the geological terminology dictionary, this application maps entities to the most matching target geological terms. Expert opinions are then used for further verification, eliminating terminological ambiguity at its source and ensuring consistent entity identification. Finally, this application incorporates the contextual information of geological entities (first related data) and their similarity to terms into the relation extraction model, and supplements the domain rules with expert opinions, enabling the model to integrate geological knowledge. This application improves the logic and accuracy of relation extraction by reasoning about relationships between entities; it automatically extracts entity and associated data, reducing manual annotation workload by more than 70%; it directly incorporates expert opinions into the model building process, supporting the rapid integration of new terms into the model through similarity matching and expert confirmation, achieving dynamic knowledge updates without model reconstruction; by linking the original data (such as text blocks containing entities), similarity calculation basis, and expert opinions throughout the process, this application ensures that every piece of knowledge (entity, relation, attribute) output by the model can be traced back to its source; at the same time, it uses a geological terminology dictionary as a benchmark to ensure that the output conforms to domain common sense, significantly improving the interpretability and credibility of the model. Attached Figure Description

[0015] To more clearly illustrate the technical solutions of the embodiments of this application, the accompanying drawings used in the embodiments of this application will be briefly introduced below. It should be understood that the following drawings only show some embodiments of this application and should not be regarded as a limitation of the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0016] Figure 1 A system architecture diagram of a method for constructing a large geological model provided in this application embodiment; Figure 2 A flowchart illustrating a method for constructing a large geological model, as provided in an embodiment of this application; Figure 3 A schematic diagram of a geological large-scale model construction device provided in an embodiment of this application; Figure 4 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Detailed Implementation

[0017] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of the embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application. Unless otherwise defined, the technical or scientific terms used in this application should have the ordinary meaning understood by those skilled in the art. The words "first," "second," and similar terms used in this application do not indicate any order, quantity, or importance, but are only used to distinguish different components. The words "comprising" or "including," etc., mean that the element or object preceding the word covers the element or object listed after the word and its equivalents, but do not exclude other elements or objects. The words "connected," "coupled," or "connected," etc., are not limited to physical or mechanical connections, but can include electrical connections, whether direct or indirect. "Up," "down," "left," "right," etc., are only used to indicate relative positional relationships. When the absolute position of the described object changes, the relative positional relationship may also change accordingly.

[0018] The method for constructing a large geological model provided in this application embodiment can be applied to... Figure 1 In the system architecture shown, such as Figure 1 As shown, the system may include: a server and a terminal; wherein, the terminal includes a visual interactive interface for acquiring geological data to be processed, and, after obtaining the second association data corresponding to the geological entities, acquiring expert opinions from geological experts on the second association data of each geological entity; the server is used to execute the geological large model construction method provided in the embodiments of this application; the server may be a physical server, or a server cluster or distributed system composed of multiple physical servers, or a cloud server providing basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, content delivery networks (CDN), and big data and artificial intelligence platforms. The terminal may be a user equipment (UE) such as a mobile phone, smartphone, laptop, digital broadcast receiver, personal digital assistant (PDA), tablet computer (PAD), handheld device, vehicle-mounted device, wearable device, computing device or other processing device connected to a wireless modem, mobile station (MS), mobile terminal, etc. The terminal and the server may be directly or indirectly connected through wired or wireless communication methods, which is not limited herein.

[0019] The preferred embodiments of this application are described below with reference to the accompanying drawings. It should be understood that the preferred embodiments described herein are for illustration and explanation only and are not intended to limit this application. Furthermore, the embodiments and features in the embodiments of this application can be combined with each other without conflict.

[0020] Figure 2 This is a flowchart illustrating a method for constructing a large geological model, as provided in an embodiment of this application. Figure 2 As shown, the method may include: Step S210: Obtain the geological data to be processed.

[0021] The methods for obtaining the geological data to be processed include: This process involves acquiring publicly available raw geological data in the field of engineering exploration. Raw geological data includes geological data (such as geological reports, core descriptions, borehole records, etc.) and engineering construction data (such as historical engineering records, design schemes, construction logs, etc.). The data types of raw geological data encompass structured data (such as borehole data, seismic data, core analysis data, and other data), semi-structured data (such as XML / JSON format geological model configuration data, WEB service interface data, and other data), and unstructured data (such as topographic vector data, image raster data, IoT monitoring data, and engineering exploration reports, etc.). For unstructured image-based raw geological data, text recognition technology is used to extract the text data from the corresponding raw geological data to obtain initial geological data. For unstructured text-based raw geological data and initial geological data, natural language processing tools are used to clean the text-based raw geological data and initial geological data to obtain the geological data to be processed.

[0022] Step S220: Perform entity identification on the geological data to obtain the first association data of multiple geological entities and various geological entities.

[0023] Specifically, the first association data includes: the dependency syntax tree of geological entities, their location, and text blocks containing the corresponding geological entities; entity recognition is performed on geological data to obtain the first association data of multiple geological entities and various geological entities, specifically including: The geological data to be processed is segmented, part-of-speech tagging is performed, and dependency parsing is performed to obtain multiple words, their part-of-speech tags, and dependency parsing trees. Text blocks containing each word are extracted from the geological data. The text blocks containing each word, the part-of-speech tags, and dependency parsing trees of each word are input into a pre-trained geological entity recognition model to obtain multiple geological entities, their dependency parsing trees, locations, and text blocks containing the corresponding geological entities.

[0024] Step S230: For any geological entity, calculate the similarity between the geological entity and the geological terms in the configured geological terminology dictionary; take the geological term with the highest similarity to the geological entity as the target geological term for the geological entity.

[0025] Specifically, the methods for constructing a geological terminology dictionary include: This project acquires correlation data for different geological terms in the field of geology, as well as metadata of academic materials containing the corresponding geological terms. The academic materials include core journals, geological expert interviews, geological literature, and geological reports. Correlation data includes: definitions, synonyms, attributes, entity types of geological entities corresponding to the geological terms, relationship types between different geological terms, and relationship weights corresponding to each relationship type. Geological terms include standard academic terms, industry slang, and emerging concepts. Standard academic terms are extracted from core journals in the field of geology; industry slang is extracted from geological expert interviews; and emerging concepts are extracted from geological literature from recent years. Entity types include: minerals, rocks (such as granite and diorite), strata (such as Cambrian), structures (such as thrust faults), and deposits (such as skarn-type deposits). Relationship types include: occurrence at, cut, overlying, and originating from. Attributes include: hardness, age, and occurrence (strike, dip, or dip direction). The correlation data for geological terms comes from geological expert annotations. For any given geological term, assign a unique term ID to it; Based on the metadata of academic materials containing the geological term, calculate the overall authority score for the geological term; the formula for calculating the overall authority score is as follows: ; in, Geological terms The overall authority rating; , , , This represents the weight coefficient of each rating dimension, satisfying... ; Geological terms Academic authority rating, , Indicates inclusion of geological terms The authority coefficient of the i-th academic document; This indicates an indicator function, which is used when the i-th academic document contains geological terms. The time is 1; Indicates inclusion of geological terms The quantity of academic materials; Geological terms The frequency of use rating is based on geological terminology. Number of times it appears in core journals, geological literature, and expert interviews; Geological terms Emerging status score ; Represents the time decay constant; Representing the current time and geological terms respectively First appearance time; Geological terms Industry usage ratio score , , They represent geological terms respectively. The frequency of use within the industry and overall was obtained through statistics from relevant publicly available academic materials; If the overall authority score of a geological term is greater than the configured authority threshold, then the geological term will be used as the target geological term. A geological terminology dictionary is constructed based on the term IDs and associated data of different target geological terms. Specifically, a synonym mapping table is constructed based on the synonyms of the target geological terms. The synonym mapping table includes non-standard expressions and standard terms for the corresponding geological terms. Non-standard expressions include abbreviations, colloquial names, misspellings, and foreign abbreviations. The geological terminology dictionary is also constructed based on the term IDs, synonym mapping tables, definitions, and attributes of different target geological terms, as well as the entity types corresponding to the respective geological terms, the relationship types between different target geological terms, and the relationship weights corresponding to the relationship types.

[0026] The similarity calculation formula is as follows: ; in, Represents geological entities Geological terminology The similarity between them; Represents geological entities Geological terminology The semantic similarity between them is calculated using word vectors or pre-trained models; Represents geological entities Geological terminology Structural similarity between characters is based on character edit distance or phonological similarity. Geological terms Frequency of use in the geological field; , , These represent the weight coefficients of the similarity scores for each item; Specifically, the geological terms with the highest similarity to the geological entity will be used as the target geological terms for the geological entity, including: If the similarity between a geological entity and the geological term with the highest similarity is not less than the configured first similarity threshold, then the geological term will be used as the target geological term for the geological entity; in this case, the target geological term is an alternative name or subordinate concept of the geological entity, etc. If the similarity between a geological entity and the geological term with the highest similarity is less than a first similarity threshold but greater than a configured second similarity threshold, then the recognition probability of the geological entity is calculated. This recognition probability, along with the similarity to the geological term with the highest similarity, is used to generate mapping information to be reviewed. This information is then used by experts to select the corresponding geological term from a geological terminology dictionary, and feedback is obtained from the experts regarding the mapping information to be reviewed. From this feedback, the selected geological term corresponding to the geological entity is extracted as the target geological term for the geological entity. At this point, the review focuses on the relationship between the geological entity and the geological data. The formula for calculating the probability of identifying geological entities is as follows:

[0027] in, Represents geological entities Recognition probability; t represents the geological entity The corresponding word; 'c' indicates the inclusion of geological entities. The text block; D represents the geological terminology dictionary and geological rule base. The geological rule base contains entity trigger words (indicating structural relationships), attribute keywords (corresponding attribute names), sentence templates (such as XX controlling the distribution of YYY deposits), as well as geological axioms (including temporal constraints such as fault cutting strata and spatial topological rules such as the inclusion relationship of ore bodies within rock strata) and Datalog inference rules. The geological entity recognition model is based on geological entities. Corresponding words, parts of speech, and geological entities The probability of recognizing geological entity e from the text block; This represents the probability of recognizing geological entity e by the geological entity recognition model based on a geological terminology dictionary and configured geological rules; The geological entity recognition model is based on the inclusion of geological entities. The probability of recognizing geological entity e from the text block; , , This indicates the identification weight coefficients, and + + = 1.

[0028] If the similarity between a geological entity and the geological term with the highest similarity is no greater than the second highest similarity, new term review information is generated for experts to determine whether the geological entity is a new term; the associated data of the geological entities determined by the experts to be new terms are obtained, and the geological term dictionary is updated.

[0029] Step S240: Input the geological entity, the first association data of the geological entity, and the similarity between the geological entity and the target geological term into the pre-trained relation extraction model to obtain the second association data corresponding to the geological entity.

[0030] Among them, the relation extraction model can not only extract the relationship between geological entities and target geological terms, but also the relationship between geological entities and geological entities as well as between geological entities and attributes, thereby constructing a knowledge graph. This can supplement the geological terminology dictionary with relevant attribute information of geological terms, and can also add new geological terms.

[0031] Specifically, relation extraction models include: The text encoding layer is used to encode text blocks containing input geological entities and the locations of the geological entities, to obtain context feature vectors and global semantic vectors of text blocks; The syntax coding layer is used to encode the dependency paths of the input geological entities and their dependency syntax trees to obtain syntax feature vectors. The rule encoding layer is used to encode the geological rules and geological terminology dictionary in the configured geological rule base to obtain rule feature vectors; The feature fusion and rule adjustment attention layer is used to concatenate the geological entity and its corresponding context feature vector, syntactic feature vector and rule feature vector based on the similarity between the geological entity and the target geological term, and then perform linear transformation and attention weighting to obtain a weighted feature vector. Specifically, based on the similarity between geological entities and target geological terms, the geological entity and its corresponding contextual feature vector, text block global semantic vector, syntactic feature vector, and rule feature vector are concatenated and then linearly transformed to obtain the rule-guided attention query vector; the attention weight of the attention query vector is calculated using the following formula: ;in, Represents the attention weights of the attention query vector; This represents the rule-guided attention query vector. ; Indicates the first The weighting coefficients of each attention factor; Indicates the first The characteristic function of each attention factor; and They represent the first The weight matrix and bias vector corresponding to each attention factor It is determined based on the similarity between the geological entity and the target geological terminology; the higher the similarity, the better. The higher; This represents the concatenated vector after a linear transformation; The relationship classification layer is used to determine the second association data and corresponding confidence level of geological entities based on the weighted feature vector. The second association data includes: the relationship type between the geological entity and other geological entities, the relationship type between the geological entity and the target geological term, and the attribute information of the geological entity.

[0032] Step S250: Obtain expert opinions from geological experts on the second correlation data of various geological entities; construct a large geological model based on the geological terminology dictionary, various geological entities and their corresponding first correlation data, second correlation data, similarity with the corresponding target geological terms and expert opinions.

[0033] Specifically, the terminal displays the second correlation data of geological entities to different geological experts through a visual interactive interface, and obtains expert opinions from different geological experts on the second correlation data of geological entities; geological experts can confirm, modify, delete and add geological entities and corresponding second correlation data.

[0034] Specifically, the methods for constructing large-scale geological models include: (1) Construct an initial geological model; based on expert opinions and the configured geological rule base, revise the second correlation data of each geological entity to obtain the third correlation data of each geological entity; (2) Based on the geological rule base, the geological terminology dictionary, various geological entities, and the corresponding first and third related data and similarity with the corresponding target geological terms, an initial geological knowledge graph is constructed; wherein, taking various geological terms and geological entities in the geological terminology dictionary as nodes, edges are established between nodes according to the geological rule base, the third related data of the corresponding geological entities, and the similarity with the corresponding target geological terms, and attribute information is added to the nodes; according to the source of the text block containing the corresponding geological entity, the attribute information of the nodes is associated with the attribute source to obtain the geological knowledge graph; (3) Using the Datalog inference engine, update the geological rule base, the initial geological knowledge graph and the geological terminology dictionary to obtain the geological knowledge graph and the updated geological rule base and geological terminology dictionary; Specifically, from the geological rule base, the first geological rule corresponding to the target geological term and the second geological rule corresponding to the relation type are matched; the Datalog inference engine is used to perform rule inference on the first and second geological rules to obtain multiple implicit geological rules; the confidence of each implicit geological rule is calculated, and the consistency between each implicit geological rule and the initial geological knowledge graph is checked; The system obtains correction opinions from geological experts based on confidence and consistency verification results for each implicit geological rule. Specifically, it graphically displays the confidence and consistency verification results of the initial geological knowledge graph and each implicit geological rule, facilitating comprehensive review by geological experts. Implicit geological rules with confidence scores below the configured confidence threshold are highlighted. The system records modifications, additions, and deletions made by geological experts to the implicit geological rules and the initial geological knowledge graph, and summarizes the opinions of different geological experts to obtain correction opinions. Based on the acquired implicit geological rules and corresponding correction suggestions, the geological rule base, the initial geological knowledge map, and the geological terminology dictionary are updated to obtain the updated geological rule base, geological knowledge map, and geological terminology dictionary. (4) Using geological knowledge graphs, the pre-constructed initial geological model is trained to obtain the geological model; Specifically, entities and relationships in the geological knowledge graph are transformed into low-dimensional vectors to generate structured samples from the geological knowledge graph; entities and relationships in the geological data to be processed are labeled with the geological knowledge graph to generate training data with semantic labels; embedding vectors from the geological knowledge graph are incorporated into the initial geological big model; the distributed training framework LLaMA-Factory is used to quickly train and update the geological big model; a prompting learning approach can be adopted, using relationships from the geological knowledge graph as prompts to guide the initial geological big model to learn domain knowledge; and the similarity between candidate entities and target geological terms is used to strengthen the association learning of highly similar entities during training to obtain the geological big model.

[0035] In another embodiment of this application, after obtaining the updated geological knowledge map, the method further includes: A dynamic quality optimization function is used to evaluate the quality of the updated geological knowledge map; the dynamic quality optimization function is as follows:

[0036] in, This represents the overall quality score of the geological knowledge graph in the (n+1)th round; This represents the quality score of the geological knowledge graph in the nth round. , and These represent the weight coefficients for expert contribution, machine learning, and penalty term, respectively. This indicates the quality improvement resulting from the nth round of revisions. =w conf *Conf improve (n)+ w cover *Cover improve (n) + w prec *Prec improve (n); This represents the quality improvement achieved in the nth round of machine learning. = Learning rate *Gradient(Loss function , Penalty(n) represents the penalty term for inconsistency between the nth round of revisions and the machine output. = Inconsistency (expert) feedback machine output );w conf , w cover and w prec Indicates the weights for improvements in confidence, coverage, and precision; Conf improve (n), Cover improve (n) and Prec improve (n) represents the improvement value of each indicator in the nth round; Learning rate Represents the machine learning rate; Gradient(·) represents the gradient of the loss function with respect to the model parameters; This represents the model parameters in the nth round; This represents the inconsistency penalty coefficient; Inconsistency represents the measure of inconsistency between expert feedback and machine output. The updated geological knowledge map was adjusted based on the quality assessment results.

[0037] Corresponding to the above method, embodiments of this application also provide a device for constructing a large geological model, such as... Figure 3 As shown, the device includes: The first acquisition unit 310 is used to acquire geological data to be processed; The identification unit 320 is used to identify entities in geological data and obtain the first association data of multiple geological entities and various geological entities. The calculation unit 330 is used to calculate the similarity between the geological entity and various geological terms in the configured geological terminology dictionary for any geological entity; and to select the geological term with the highest similarity to the geological entity as the target geological term for the geological entity. Input unit 340 is used to input geological entities, first association data of geological entities, and similarity between geological entities and target geological terms into a pre-trained relation extraction model to obtain second association data corresponding to geological entities; The second acquisition unit 350 is used to acquire expert opinions from geological experts on the second correlation data of various geological entities; The 360-unit construction module is used to build a large geological model based on a geological terminology dictionary, various geological entities and their corresponding first and second related data, similarity with the corresponding target geological terms, and expert opinions.

[0038] The functions of each functional unit in the geological large-scale model construction device provided in the above embodiments of this application can be realized through the above-described method steps. Therefore, the specific working process and beneficial effects of each unit in the geological large-scale model construction device provided in the embodiments of this application will not be repeated here.

[0039] This application also provides an electronic device, such as... Figure 4 As shown, it includes a processor 410, a communication interface 420, a memory 430, and a communication bus 440, wherein the processor 410, the communication interface 420, and the memory 430 communicate with each other through the communication bus 440.

[0040] Memory 430 is used to store computer programs; When the processor 410 executes the program stored in the memory 430, it performs the following steps: Obtain the geological data to be processed; Entity identification is performed on geological data to obtain primary correlation data between multiple geological entities and various geological entities. For any given geological entity, calculate the similarity between the geological entity and the geological terms in the configured geological terminology dictionary. The geological term with the highest similarity to the geological entity is used as the target geological term for the geological entity. The geological entity, the first association data of the geological entity, and the similarity between the geological entity and the target geological term are input into a pre-trained relation extraction model to obtain the second association data corresponding to the geological entity. Obtain expert opinions from geological experts on the second correlation data of various geological entities; A large geological model is constructed based on a geological terminology dictionary, various geological entities and their corresponding first and second association data, similarity with the corresponding target geological terms, and expert opinions.

[0041] The communication bus mentioned above can be a Peripheral Component Interconnect (PCI) bus or an Extended Industry Standard Architecture (EISA) bus, etc. This communication bus can be divided into address bus, data bus, control bus, etc. For ease of illustration, only one thick line is used to represent it in the diagram, but this does not mean that there is only one bus or one type of bus.

[0042] The communication interface is used for communication between the aforementioned electronic devices and other devices.

[0043] The memory may include random access memory (RAM) or non-volatile memory (NVM), such as at least one disk storage device. Optionally, the memory may also be at least one storage device located remotely from the aforementioned processor.

[0044] The processors mentioned above can be general-purpose processors, including central processing units (CPUs), network processors (NPs), etc.; they can also be digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components.

[0045] The implementation methods and beneficial effects of the various components of the electronic device in the above embodiments for solving the problem can be found in [reference needed]. Figure 2 The steps in the illustrated embodiments are used to implement the electronic device. Therefore, the specific working process and beneficial effects of the electronic device provided in this application will not be repeated here.

[0046] In another embodiment provided in this application, a computer-readable storage medium is also provided, which stores instructions that, when executed on a computer, cause the computer to perform the method for constructing a large geological model as described in any of the above embodiments.

[0047] In another embodiment provided in this application, a computer program product containing instructions is also provided, which, when run on a computer, causes the computer to execute the method for constructing a geological large model as described in any of the above embodiments.

[0048] Those skilled in the art will understand that the embodiments in this application can be provided as methods, systems, or computer program products. Therefore, the embodiments in this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the embodiments in this application can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0049] This application describes embodiments of methods, apparatus (systems), and computer program products according to embodiments of this application with reference to flowchart illustrations and / or block diagrams. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0050] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0051] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0052] Although preferred embodiments have been described in this application, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments as well as all changes and modifications falling within the scope of the embodiments of this application.

[0053] Obviously, those skilled in the art can make various modifications and variations to the embodiments of this application without departing from the spirit and scope of the embodiments of this application. Therefore, if these modifications and variations to the embodiments of this application fall within the scope of this application and its equivalents, then these modifications and variations are also intended to be included in the embodiments of this application.

Claims

1. A method for constructing a large geological model, characterized in that, The method includes: Obtain the geological data to be processed; Entity recognition is performed on the geological data to obtain multiple geological entities and first association data of each geological entity; the first association data includes: the dependency syntax tree of the geological entity, its location, and the text block containing the corresponding geological entity; For any given geological entity, calculate the similarity between the geological entity and the geological terms in the configured geological terminology dictionary. The geological term with the highest similarity to the geological entity is taken as the target geological term for the geological entity. The geological entity, the first association data of the geological entity, and the similarity between the geological entity and the target geological term are input into a pre-trained relation extraction model to obtain the second association data corresponding to the geological entity. The second association data includes: the relationship type between the geological entity and other geological entities besides the geological entity, the relationship type between the geological entity and the target geological term, and the attribute information of the geological entity. Obtain expert opinions from geological experts on the second correlation data of various geological entities; Based on the geological terminology dictionary, various geological entities and their corresponding first and second association data, similarity with corresponding target geological terms, and expert opinions, a large-scale geological model is constructed, including: revising the second association data of various geological entities according to the expert opinions and a configured geological rule base to obtain the third association data of each geological entity; constructing an initial geological knowledge graph based on the geological rule base, the geological terminology dictionary, various geological entities and their corresponding first and third association data, and similarity with corresponding target geological terms; updating the geological rule base, the initial geological knowledge graph, and the geological terminology dictionary using the Datalog inference engine to obtain the geological knowledge graph and the updated geological rule base and geological terminology dictionary; and training the pre-constructed initial large-scale geological model using the geological knowledge graph to obtain the large-scale geological model.

2. The method as described in claim 1, characterized in that, The configuration method of the geological terminology dictionary includes: To obtain the associated data of different geological terms in the field of geology, as well as the metadata of academic materials containing the corresponding geological terms; For any given geological term, assign a unique term ID to that geological term; Calculate the overall authority score of the geological term based on the metadata of academic materials containing the geological term; If the overall authority score of the geological term is greater than the configured authority threshold, then the geological term will be used as the target geological term. A geological terminology dictionary is constructed based on the term IDs and associated data of different target geological terms.

3. The method as described in claim 1, characterized in that, Entity identification was performed on the geological data to obtain multiple geological entities, including: The geological data to be processed is segmented, part-of-speech tagging is performed, and dependency parsing is performed to obtain multiple words, the part-of-speech tags of the corresponding words, and dependency parsing trees. Extract text blocks containing each word from the geological data; The text blocks containing each word, the part-of-speech tag and dependency syntax tree of each word are input into the pre-trained geological entity recognition model to obtain multiple geological entities and their corresponding dependency syntax trees, locations and text blocks containing the corresponding geological entities.

4. The method as described in claim 3, characterized in that, The relation extraction model includes: The text encoding layer is used to encode text blocks containing input geological entities and the locations of the geological entities to obtain context feature vectors and global semantic vectors of text blocks. The syntactic coding layer is used to encode the dependency paths of the input geological entities and their dependency syntax trees to obtain syntactic feature vectors. The rule encoding layer is used to encode the geological rules and geological terminology dictionary in the configured geological rule base to obtain rule feature vectors; The feature fusion and rule adjustment attention layer is used to concatenate the geological entity and its corresponding context feature vector, text block global semantic vector, syntactic feature vector and rule feature vector based on the similarity between the geological entity and the target geological term, and then perform linear transformation and attention weighting to obtain a weighted feature vector. The relation classification layer is used to determine the second association data and corresponding confidence level of the geological entity based on the weighted feature vector.

5. The method as described in claim 1, characterized in that, Based on the geological rule base, the geological terminology dictionary, various geological entities, and corresponding first and third association data, as well as the similarity with corresponding target geological terms, a geological knowledge graph is constructed, including: Using geological terms and entities from the geological terminology dictionary as nodes, edges are established between nodes based on the geological rule base, the third-party associated data of the corresponding geological entities, and the similarity with the corresponding target geological terms. Attribute information is added to the nodes. Based on the source of the text blocks containing the corresponding geological entities, the attribute information of the nodes is associated with the attribute source to obtain a geological knowledge graph.

6. The method as described in claim 1, characterized in that, Updating the geological rule base, the initial geological knowledge graph, and the geological terminology dictionary includes: From the geological rule base, match the first geological rule corresponding to the target geological term and the second geological rule corresponding to the relationship type; The Datalog inference engine is used to perform rule inference on the first geological rule and the second geological rule to obtain multiple implicit geological rules; Calculate the confidence level of each implicit geological rule, and perform consistency verification between each implicit geological rule and the initial geological knowledge graph; Obtain the modification opinions of geological experts on each implicit geological rule based on the confidence level and consistency verification results; Based on the acquired implicit geological rules and corresponding correction suggestions, the geological rule base, the initial geological knowledge graph, and the geological terminology dictionary are updated to obtain the updated geological rule base, geological knowledge graph, and geological terminology dictionary.

7. A device for constructing a large geological model, characterized in that, The device includes: The first acquisition unit is used to acquire geological data to be processed. The identification unit is used to perform entity identification on the geological data to obtain multiple geological entities and first association data of each geological entity; the first association data includes: the dependency syntax tree of the geological entity, its location, and the text block containing the corresponding geological entity; The calculation unit is used to calculate the similarity between any geological entity and various geological terms in a configured geological terminology dictionary for any given geological entity; and to select the geological term with the highest similarity to the geological entity as the target geological term for the geological entity. The input unit is used to input the geological entity, the first association data of the geological entity, and the similarity between the geological entity and the target geological term into a pre-trained relation extraction model to obtain the second association data corresponding to the geological entity; the second association data includes: the relationship type between the geological entity and other geological entities besides the geological entity, the relationship type between the geological entity and the target geological term, and the attribute information of the geological entity; The second acquisition unit is used to acquire expert opinions from geological experts on the second correlation data of various geological entities; The construction unit is used to construct a large-scale geological model based on the geological terminology dictionary, various geological entities, and corresponding first and second association data, similarity with corresponding target geological terms, and expert opinions. This includes: correcting the second association data of various geological entities according to the expert opinions and a configured geological rule base to obtain third association data for each geological entity; constructing an initial geological knowledge graph based on the geological rule base, the geological terminology dictionary, various geological entities, and corresponding first and third association data, and similarity with corresponding target geological terms; updating the geological rule base, the initial geological knowledge graph, and the geological terminology dictionary using the Datalog inference engine to obtain the geological knowledge graph and the updated geological rule base and geological terminology dictionary; and training the pre-constructed initial large-scale geological model using the geological knowledge graph to obtain the large-scale geological model.

8. An electronic device, characterized in that, The electronic device includes a processor, a communication interface, a memory, and a communication bus, wherein the processor, the communication interface, and the memory communicate with each other through the communication bus; Memory, used to store computer programs; A processor, when executing a program stored in memory, implements the method of any one of claims 1-6.

9. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, implements the method described in any one of claims 1-6.

Images