Geological prospecting literature knowledge extraction method and system based on large model

Abstract

The invention discloses a geological prospecting literature knowledge extraction method and system based on a large model, and the method comprises the steps: building a triple learning model, and training the triple learning model to recognize triple data in geological prospecting literature and the existence form of the triple data in the geological prospecting literature; collecting geological prospecting literatures from various literature entrances by taking the triple learning model as an engine, storing the collected geological prospecting literatures into a preset database, and carrying out classification labeling work based on the geological prospecting literatures; extracting triple data from the geological prospecting literature stored in the database by using the triple learning model, and forming leaf-shaped knowledge points based on the extracted triple data and storing the leaf-shaped knowledge points into a knowledge base; organizing the leafy knowledge points through business logic to form a tree-shaped or star-shaped knowledge structure; and drawing a knowledge graph based on the knowledge structure.

Description

Technical Field

[0001] The present invention relates to the technical field of machine learning, and in particular to a method and system for extracting geological prospecting literature knowledge based on a large model. Background Art

[0002] With the advancement of science and technology and the development of business, new concepts and terms will continue to appear in newly generated documents. Traditional manual reading of documents to extract information will make it impossible to understand the documents due to problems such as knowledge reserves. In addition, since the amount of knowledge in the field of prospecting and exploration is large and the professional knowledge required for different approach angles is also different, different documents require the assistance of experts in different sub-fields to be realized. Therefore, the existing mineral document extraction work is completed by manpower. At the same time, due to the large amount of mineral-related documents and relatively long length, manual extraction requires complete reading of the documents and identification and extraction. The whole process is very time-consuming and inefficient.

[0003] On the other hand, in the process of reading and recognition, triple data is usually introduced. Triple data is a basic data structure composed of three elements: subject (entity), predicate (relationship) and object (entity). This structure is used to describe the relationship between things, usually expressed as (subject, predicate, object). Subject and object are entities, which can be concrete things or abstract concepts. Predicate represents the relationship between subject and object. In the process of reading existing literature and identifying triples, some professional content is incomprehensible to ordinary technical workers, which easily leads to some difficulties in extracting triples or not knowing how to extract them. It is necessary to rely on professionals or experts to assist in reading. If the support of professionals or experts cannot be obtained, the document information cannot be extracted. Manual reading of literature and identification and extraction of triple information are limited by human energy and fatigue, and errors are more likely to occur in the reading and recognition process than machines.

[0004] Mineral research and prospecting research have gone through several stages in the course of development and will continue to move forward in the future. In this process, the language structure and writing logic of the documents written by researchers are very different. Readers and extractors can only recognize part of the language structure and writing logic. They have difficulty in reading the language structure and writing logic used in ancient and rare documents and documents that have not yet been produced in the future.

[0005] Manual reading is very inefficient when faced with multilingual and complex language document environments, and may even be unable to complete reading and extraction tasks. Summary of the invention

[0006] In order to solve the above problems, the purpose of the present invention is to provide a geological prospecting literature knowledge extraction method and system based on a large model, which replaces manual work by taking advantage of the fast reading and precise recognition of the large model to realize the extraction of mineral literature triples.

[0007] The present invention provides a method for extracting geological prospecting literature knowledge based on a large model, the method comprising: Establishing a triplet learning model, and training the triplet learning model to identify triplet data in geological prospecting literature and the form in which the triplet data exists in the geological prospecting literature; Using the triple learning model as an engine, geological prospecting documents are collected from various document portals, the collected geological prospecting documents are stored in a pre-set database, and classification and annotation work is performed based on the geological prospecting documents; Extracting triplet data from geological prospecting literature stored in the database using the triplet learning model, and forming leaf-shaped knowledge points based on the extracted triplet data and storing them in a knowledge base; Organizing the leaf-shaped knowledge points through business logic to form a tree-shaped or star-shaped knowledge structure; A knowledge graph is drawn based on the knowledge structure.

[0008] Optionally, the classification and annotation work based on the geological prospecting literature includes: Label the names of geological entities; Mark the attribute information of geological entities; Annotate the mutual relationships between geological entities and the semantic relationships between geological entities and attribute information; The article viewpoints, values ​​and methods of the geological prospecting literature are marked.

[0009] Optionally, the attribute information includes ore deposits, attribute characteristics, ore-controlling factors, and ore-prospecting marks; The deposit includes one or more of the following: environment / ore-forming environment / geological environment / tectonic domain, geological phenomenon, method / geological method / geological technology / geological principle / step; The attribute characteristics include one or more of the following: formation age / time / metallogenic age, diagenetic age, crystallization age, grade / element grade, ore body morphology, ore vein / ore body, ore section / ore belt, mining area / ore field / ore cluster, region / location, rock type, ore deposit type, resource volume / output; The prospecting criteria include one or more of the following: geophysical anomaly / element anomaly, geochemical anomaly / element anomaly, remote sensing interpretation anomaly / element anomaly; The ore-controlling factors include one or more of the following: geological body / rock body / three major rock bodies / surrounding rock type, stratum, geological event / geological process, alteration type, mineralization stage / mineralization type, mineralization / mineralization, and structure.

[0010] Optionally, extracting data using the triple learning model includes: constructing a structured promt using promt prompt word engineering; the structured promt includes: role, goal, {context, restriction}, {skill, tool}, {input rule, output rule}, and output example.

[0011] The present invention also provides a geological prospecting literature knowledge extraction system based on a large model, the system comprising one or more processors and a non-temporary computer-readable storage medium storing program instructions, when the one or more processors execute the program instructions, the one or more processors are used to implement any of the geological prospecting literature knowledge extraction methods based on a large model described above.

[0012] Optionally, the system includes: a mineral document collection module, a document information extraction and research module, an information knowledge refinement module, and a document knowledge product production module; The mineral document collection module is used to collect mineral documents through multiple channels; The literature information extraction research and development module is used to generate a literature information extraction tool to extract information based on the collected mineral literature; The information knowledge processing module is used to generate a prospecting prediction map; The literature knowledge product production module is used to provide mineral prospecting-themed service applications.

[0013] Optionally, the mineral document collection module collects mineral documents through commercial collection, independent construction, open access and / or shared exchange.

[0014] The present invention also provides a computing device, which includes any of the large model-based geological prospecting literature knowledge extraction systems described above.

[0015] The present invention also provides a computer-readable storage medium, which is used to store program code, and the program code is used to execute any of the above-mentioned methods for extracting geological prospecting literature knowledge based on a large model.

[0016] According to the present invention, since the knowledge extraction method is performed through a large model, it does not only rely on the integration of entities, relationships and attributes in structured data, semi-structured data and unstructured data, but in addition to identifying triple data, it also identifies the forms in which triples exist, including text form, hypertext form, cross-text form, image form, data form, cross-data form, etc. On the other hand, according to the present invention, not only the entity relationship attributes are annotated, but also the article views, values ​​and methods are annotated, which can ensure the generation efficiency and accuracy of the knowledge graph.

[0017] In the present invention, an entity disambiguation mechanism is established. The mechanism will first learn a large number of template documents to establish a series of entity object clustering sets, and then form a semantic dictionary pointing to the object based on the context of the target document. The dictionary is aggregated under the established object clustering. When constructing triples, a secondary tracking strategy of first matching the entity object clustering and then matching the semantic dictionary is used to achieve entity disambiguation.

[0018] The geological prospecting literature knowledge extraction system based on the large model of the present invention has accumulated more professional knowledge through learning with the large model, and can accurately identify the information that needs to be extracted in the professional content, reducing the dependence on external assistance. By adopting machine reading and machine extraction to replace manual work, the error rate caused by fatigue in the reading and extraction process is eliminated. Further, the large model is updated and upgraded through machine learning, so that the large model can always understand the language structure and writing logic of each document, and accurately extract the required information. To this end, the present invention establishes a model iteration mechanism, and the mechanism includes steps such as global new reading, prompt word engineering, new structure identification and new prompt word engineering iteration. The language structure and prompt word engineering in the new document are obtained through global new reading, and the similarity is matched with the existing model. When it is lower than the set threshold, the language structure extracted by reading the new document is set as a new structure and enters the structure library, and a secondary prompt word engineering iteration model is established, wherein the original prompt word engineering is corrected as a primary iteration, and the prompt word engineering model library is updated as a secondary iteration. Taking advantage of the large model's cross-language environment, multilingual and complex language documents are directly read by the large model without translation, and effective information is extracted. Through pre-training, we can establish the ability to identify special corpus, and through multi-sample exercises, we can form the disassembly and entity extraction of documents in special corpus environments. The system can be configured with all regular corpus documents, special corpus environment documents, and mixed corpus documents, so that the large language model can call different rules for extraction.

[0019] In the present invention, in order to effectively find errors in knowledge graph data and improve data quality, the system uses a deep learning mechanism to continuously optimize the large language model for extracting triples, establish a conventional document prompt word project and a comparison document prompt word project. When the output result is significantly different from the expected result, the deep learning mechanism is activated to disassemble the manually extracted triple sample step by step, and understand and learn step by step.

[0020] Iteratively learn the principles of entity extraction from long-distance context information, and use the existing samples and new samples after manual error correction to bidirectionally train the model, so that the prompt word engineering can intelligently judge the impact of long-distance context on triples and accurately perform group matching.

[0021] Based on the following detailed description of specific embodiments of the present invention in conjunction with the accompanying drawings, those skilled in the art will become more aware of the above and other objects, advantages and features of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] Various other advantages and benefits will become apparent to those of ordinary skill in the art by reading the detailed description of the preferred embodiments below. The accompanying drawings are only for the purpose of illustrating the preferred embodiments and are not to be considered as limiting the present invention. Moreover, the same reference symbols are used throughout the accompanying drawings to represent the same components. In the accompanying drawings: Figure 1 It is a schematic diagram of the flow of a method for extracting geological prospecting literature knowledge based on a large model according to an embodiment of the present invention; Figure 2 is a schematic diagram of a triplet learning model according to an embodiment of the present invention; Figure 3 It is a schematic diagram of a geological prospecting literature knowledge extraction system based on a large model according to an embodiment of the present invention; Figure 4 It is a schematic diagram of a geological prospecting literature knowledge extraction system based on a large model according to another embodiment of the present invention. DETAILED DESCRIPTION

[0023] The embodiments of the present invention are described below in conjunction with the accompanying drawings. Those skilled in the art should understand that these embodiments are only used to explain the present invention and are not limiting.

[0024] The embodiment of the present invention provides a method for extracting geological prospecting literature knowledge based on a large model, such as Figure 1 As shown, the large-model-based geological prospecting literature knowledge extraction method of an embodiment of the present invention may include the following steps S1 to S6.

[0025] S1, establishing a triplet learning model, and training the triplet learning model to identify triplet data in geological prospecting literature and the form in which the triplet data exists in the geological prospecting literature.

[0026] Triple data is a basic data structure consisting of three elements: subject (entity), predicate (relationship) and object (entity). This structure is used to describe the relationship between things, usually expressed as (subject, predicate, object). The subject and object are entities, which can be concrete things or abstract concepts, and the predicate represents the relationship between the subject and the object.

[0027] The triple learning model mentioned in this embodiment is a machine learning model, which can be used to identify triple data in geological prospecting documents and the forms in which triple data exist in geological prospecting documents. Triple data can include subject entities, relationships, and object entities, and the forms in which triple data exist in geological prospecting documents can include text form, hypertext form, cross-text form, image form, data form, cross-data form, etc.

[0028] S2, using the triple learning model as an engine to collect geological prospecting literature from various literature portals; wherein the collection condition is to include identified prospecting-related triple data.

[0029] S3, storing the collected geological prospecting documents into a pre-set database, and performing classification and annotation work based on the geological prospecting documents.

[0030] S4, extracting triplet data from geological prospecting literature stored in the database using the triplet learning model, and forming leaf-shaped knowledge points based on the extracted triplet data and storing them in a knowledge base.

[0031] S5, organize the leaf-shaped knowledge points to form a tree-shaped or star-shaped knowledge structure. Leaf-shaped structure is an important concept in mathematics, which is mainly used to describe the hierarchical structure on the manifold. For the leaf-shaped knowledge points formed, the leaf-shaped knowledge points can be organized through business logic to form a tree-shaped or star-shaped knowledge structure. A tree structure usually represents a hierarchical relationship, while a star structure may represent the relationship between a central entity and multiple related entities. When constructing a tree-shaped knowledge structure, you can first determine the root node: select a core entity as the root node, which is usually the broadest and most abstract concept; then determine the branch: starting from the root node, add child nodes according to the business logic, each child node represents a more specific entity or concept; then determine the leaf node: the bottom-level node, representing a specific knowledge point or data.

[0032] When building a star structure, you can first determine the central node: select a core entity as the central node, which is usually the most important entity in the business logic; then determine the radial nodes: starting from the central node, connect all related entities, and each connection represents a knowledge point.

[0033] S6, drawing a knowledge graph based on the knowledge structure.

[0034] The method for extracting knowledge from geological prospecting literature based on a large model in an embodiment of the present invention replaces manual work by taking advantage of the large model's rapid reading and accurate recognition to achieve the extraction of triple data from mineral literature. Since the triple learning model has accumulated a lot of professional knowledge through learning, it can accurately identify the information that needs to be extracted from professional content, reducing the reliance on external assistance. By replacing manual work with machine reading and machine extraction, the error rate caused by fatigue in the reading and extraction process is eliminated. At the same time, the advantage of the large model's cross-language environment can be utilized. Without translation, the large model can directly read multilingual and complex language literature to extract effective information.

[0035] In an optional embodiment of the present invention, when drawing a knowledge graph for knowledge fusion, the ambiguity between entities, relationships, attributes and other referents and factual objects is eliminated, so that knowledge from different sources can be standardized and integrated. Therefore, an entity disambiguation mechanism can be established. First, a large number of template documents are learned to establish a series of entity object clustering sets, and then a semantic dictionary pointing to the object is formed based on the context of the target document. The dictionary is aggregated under the established object cluster. When constructing triples, a secondary tracking strategy of matching the entity object cluster first and then matching the semantic dictionary is used to achieve entity disambiguation.

[0036] Optionally, the knowledge graph is not generated in one go, but is a slowly accumulated process that requires continuous updating and iteration. Therefore, a model iteration mechanism can also be established, which includes steps such as global new reading, prompt word engineering, new structure identification, and new prompt word engineering iteration. The language structure and prompt word engineering in the new document are obtained through global new reading, and the similarity is matched with the existing model. When it is lower than the set threshold, the language structure extracted from reading the new document is set as the new structure and enters the structure library, and a secondary prompt word engineering iteration model is established, in which the correction of the original prompt word engineering is the first iteration, and the updating of the prompt word engineering model library is the second iteration.

[0037] Furthermore, the deep learning mechanism can be used to continuously optimize the triple extraction learning model, establish a conventional literature prompt word project and a comparison literature prompt word project. When the output result is significantly different from the expected, the deep learning mechanism is activated to decompose the steps of the manually extracted triple sample, and understand and learn step by step. Iteratively learn the principles of entity extraction in long-distance context information, use existing samples and new samples after manual error correction to train the model in a two-way manner, so that the prompt word project can intelligently judge the impact of long-distance context on triples, and accurately perform group matching, so as to effectively detect errors in knowledge graph data and improve data quality.

[0038] In an optional embodiment of the present invention, the triple information in the document should be annotated before extraction, that is, the classification and annotation work based on the geological prospecting document includes: (1) Label the names of geological entities; (2) Labeling the attribute information of geological entities; (3) Annotate the relationships between geological entities and the semantic relationships between geological entities and attribute information; (4) Annotate the viewpoints, values ​​and methods of the articles in the geological prospecting literature.

[0039] Combination Figure 2 It can be seen that the attribute information includes ore deposits, attribute characteristics, ore-controlling factors, and prospecting signs.

[0040] The mineral deposit includes one or more of the following: environment / ore-forming environment / geological environment / tectonic domain, geological phenomenon, method / geological method / geological technology / geological principle / step.

[0041] The attribute characteristics include one or more of the following: formation age / time / metallogenic age, diagenetic age, crystallization age, grade / element grade, ore body morphology, vein / ore body, ore section / ore belt, mining area / ore field / ore cluster, region / location, rock type, ore deposit type, resource quantity / output.

[0042] The prospecting standards include one or more of the following: geophysical anomalies / element anomalies, geochemical anomalies / element anomalies, and remote sensing interpretation anomalies / element anomalies.

[0043] The ore-controlling factors include one or more of the following: geological body / rock body / three major rock bodies / surrounding rock type, stratum, geological event / geological process, alteration type, mineralization stage / mineralization type, mineralization / mineralization, and structure.

[0044] In an optional embodiment of the present invention, when using the triple learning model for knowledge extraction, a structured promt can be constructed using promt prompt word engineering; the structured promt includes: role, goal, {context, restriction}, {skill, tool}, {input rule, output rule}, and output example. The construction example is as follows: Role: You are a professional geological prospector and gold mineralogist, and I will input a document to you.

[0045] Target: (1) Accurately and efficiently extract various entities closely related to gold prospecting from the given literature paragraphs, and clearly extract the specific relationships between the entities.

[0046] (2) Strictly organize the extracted results into the table headers of "serial number, head entity type, head entity, relationship between entities, tail entity type, tail entity, relationship attribute, other attributes, original sentence, article number, and remarks". Carefully draw an accurate and clear table, and be careful not to repeatedly extract the same or similar triples of the same sentence.

[0047] Output logic: (3) The extraction logic is: first separate each natural paragraph of the document; then extract triples from each natural paragraph and output the results in batches.

[0048] (4) The extracted entities must comprehensively cover the key entities and their relationships mentioned in the document to build a complete relationship network.

[0049] The identified triple data can be filled into the table according to the table relationship. The table format is shown in the table below.

[0050] Table 1 Serial number Header Entity Type Head Entity Relationships between entities Tail Entity Type Tail Entity Relationship properties Other properties Original paragraph Article number Among them, the head entity type can include 26 types such as region (administrative division, geological body, plateau, spraying), mining area, mining belt, ore deposit, stratum, rock mass, etc.; The relationships between entities can include 27 types of relationships, such as belonging to, located in, developed, experienced, related, controlled, displayed, and predicted.

[0051] The tail entity types can include 26 entity types such as regions (administrative divisions, geological bodies, plateaus, basins), mining areas, mining belts, ore deposits, ore veins, strata, and rock masses.

[0052] Relationship attributes may include: facts, opinions, and methods.

[0053] Other attributes may include: person, time, place, value.

[0054] Finally, the reviewed content can be stored in the database, such as Table 2.

[0055] Table 2 Serial number Header Entity Type Head Entity Relationships between entities Tail Entity Type Tail Entity Relationship properties Other properties Original paragraph Article number 1 Region (geological body) South Qinling tectonic belt Formed in Geological events Indosinian-Yanshanian Qinling orogeny fact Place The South Qinling tectonic belt is a composite orogenic belt formed by the superposition of multiple tectonic movements. 42 2 Mineral Deposits Rich gold mines belong Rock mass Biotite monzonitic granite fact Place The rich gold deposits in a certain county are produced in the Dongjiangkou rock body in the southern Qinling tectonic belt. 42 3 Mineral Deposits Rich gold mines Formed in Mineralization Age About 200 Ma fact time The deposit was formed in the post-Pengzheng tectonic environment during the Qinling orogeny during the Indosinian-Yanshanian period, when the tectonic structure changed from collision to extension. 42 4 Deposit Type Rich gold mines belong Mineralization type Sulphide quartz vein type Viewpoint Deposit type: tailing sulfide quartz vein type and broken zone altered rock type 42 5 Geological phenomenon Occurrence of gold have feature Encapsulated gold, fissure gold and interstitial gold fact Gold can be found in three states: inclusion gold, fissure gold, and interstitial gold. 42 6 Geology (geological bodies) Qinling tectonic belt yes Region (Administrative Division) A country fact Place The Qinling tectonic belt is one of the main gold mineralization belts in a certain country. 42 7 Mineral Deposits Rich gold mines belong Deposit size Medium fact The deposit is medium-sized 42 8 Ore body Gold ore body lie in structure Gently inclined fracture fact The ore bodies are mainly located in the gently inclined faults developed in the rock mass. 42 9 Ore type ore include type Sulphide quartz vein type and fracture zone altered rock type Viewpoint Ore types can be divided into sulfide quartz vein type and broken zone altered rock type 42 10 Geological body Dongjiangkou Rock Mass Formed in Geological events Post-collision structural environment fact Place The Dongjiangkou rock mass was formed in a typical post-collision tectonic environment. 42 11 Geological events Magmatic activity Formed in Geological events Indosinian Period fact time In recent years, Mousheng Mining Development Industry and Trade Company has discovered rich gold mines in the southern Qinling area. 42 12 Mineral Deposits Rich gold mines Formed in Geological body Dongjiangkou Rock Mass fact Place Rich gold deposits are found in the Dongjiangkou rock body 42 13 Geological body Dongjiangkou Rock Mass include Rock mass Gaoqiaojie, Dongjiangkou, Zhashui, Pingganchuan rock mass fact Place The Dongjiangkou rock mass group includes four geologically unconnected rock masses to the east. 42 14 Geological body Dongjiangkou Rock Mass Formed in Geological events Between 220 and 200 Ma fact time The Dongjiangkou rock mass was formed during the subduction, collision, exhumation and extension of the Yangtze Plate and the North China Plate to form post-collision granite. 42 15 Geological structure fracture control Geological phenomenon Spatial distribution characteristics of shallow-ultra-shallow intrusions fact Faults control the spatial distribution characteristics of shallow-ultra-shallow intrusions 42 The embodiment of the present invention also provides a geological prospecting literature knowledge extraction system based on a large model, which is used to execute the geological prospecting literature knowledge extraction method based on a large model of the above embodiment.

[0056] like Figure 3~Figure 4 As shown, the geological prospecting literature knowledge extraction system based on a large model of an embodiment of the present invention includes: a mineral literature collection module, a literature information extraction and research and development module, an information knowledge refinement module and a literature knowledge product production module.

[0057] Among them, the mineral literature collection module is used to collect mineral literature through various channels; the mineral literature collection module collects mineral literature commercially, independently, through open access and / or through shared exchange. The literature information extraction research and development module is used to generate literature information extraction tools to extract information based on the collected mineral literature; the information knowledge processing module is used to generate prospecting prediction maps; the literature knowledge product production module is used to provide prospecting-themed service applications.

[0058] An embodiment of the present invention further provides a computing device, which includes the large model-based geological prospecting literature knowledge extraction system described in the above embodiment.

[0059] An embodiment of the present invention also provides a computer-readable storage medium, which is used to store program code, and the program code is used to execute the large model-based geological prospecting literature knowledge extraction method described in the above embodiment.

[0060] The above is only a preferred embodiment of the present invention, and the protection scope of the present invention is not limited to the above embodiments. All technical solutions under the concept of the present invention belong to the protection scope of the present invention. It should be pointed out that for ordinary technicians in this technical field, some improvements and modifications without departing from the principle of the present invention should also be regarded as the protection scope of the present invention.

Claims

1. A method for extracting geological prospecting literature knowledge based on a large model, characterized in that: The method comprises: Establishing a triplet learning model, and training the triplet learning model to identify triplet data in geological prospecting literature and the form in which the triplet data exists in the geological prospecting literature; Using the triple learning model as an engine, collecting geological prospecting literature from various literature portals; The collected geological prospecting documents are stored in a pre-set database, and classification and annotation work is performed based on the geological prospecting documents; Extracting triplet data from geological prospecting literature stored in the database using the triplet learning model, and forming leaf-shaped knowledge points based on the extracted triplet data and storing them in a knowledge base; Organizing the leaf-shaped knowledge points to form a tree-shaped or star-shaped knowledge structure; A knowledge graph is drawn based on the knowledge structure.

2. The method according to claim 1, characterized in that The classification and annotation work based on the geological prospecting literature includes: Label the names of geological entities; Mark the attribute information of geological entities; Annotate the mutual relationships between geological entities and the semantic relationships between geological entities and attribute information; The article viewpoints, values ​​and methods of the geological prospecting literature are marked.

3. The method according to claim 2, characterized in that The attribute information includes mineral deposits, attribute characteristics, mineral control factors, and prospecting signs; The deposit includes one or more of the following: environment / ore-forming environment / geological environment / tectonic domain, geological phenomenon, method / geological method / geological technology / geological principle / step; The attribute characteristics include one or more of the following: formation age / time / metallogenic age, diagenetic age, crystallization age, grade / element grade, ore body morphology, ore vein / ore body, ore section / ore belt, mining area / ore field / ore cluster, region / location, rock type, ore deposit type, resource volume / output; The prospecting criteria include one or more of the following: geophysical anomaly / element anomaly, geochemical anomaly / element anomaly, remote sensing interpretation anomaly / element anomaly; The ore-controlling factors include one or more of the following: geological body / rock body / three major rock bodies / surrounding rock type, stratum, geological event / geological process, alteration type, mineralization stage / mineralization type, mineralization / mineralization, and structure.

4. The method according to claim 1, characterized in that: The triple learning model has an entity disambiguation mechanism; The entity disambiguation mechanism first learns a large number of template documents to establish a series of entity object clustering sets, and then forms a semantic dictionary pointing to the object based on the context of the target document, and aggregates the dictionary into the established object cluster. When constructing triples, a secondary tracking strategy of first matching the entity object cluster and then matching the semantic dictionary is used to achieve entity disambiguation.

5. The method according to claim 1, characterized in that The triple learning model has a model iteration mechanism; The model iteration mechanism includes the steps of global new reading, prompt word engineering, new structure recognition and new prompt word engineering iteration; The language structure and prompt word engineering in the new document are obtained through global new reading, and the similarity is matched with the existing model. When it is lower than the set threshold, the language structure extracted from reading the new document is set as the new structure and enters the structure library, and a secondary prompt word engineering iteration model is established, among which the correction of the original prompt word engineering is the first iteration, and the updating of the prompt word engineering model library is the second iteration.

6. A geological prospecting literature knowledge extraction system based on a large model, characterized in that: The system includes one or more processors and a non-transitory computer-readable storage medium storing program instructions. When the one or more processors execute the program instructions, the one or more processors are used to implement the large-model-based geological prospecting literature knowledge extraction method described in any one of claims 1-5.

7. The system according to claim 6, characterized in that The system includes: a mineral document collection module, a document information extraction and research module, an information knowledge refinement module, and a document knowledge product production module; The mineral document collection module is used to collect mineral documents through multiple channels; The literature information extraction research and development module is used to generate a literature information extraction tool to extract information based on the collected mineral literature; The information knowledge processing module is used to generate a prospecting prediction map; The literature knowledge product production module is used to provide mineral prospecting-themed service applications.

8. The system according to claim 7, characterized in that The mineral document collection module collects mineral documents through commercial collection, independent construction, open access and / or shared exchange.

9. A computing device, characterized in that The computing device includes the large model-based geological prospecting literature knowledge extraction system as described in any one of claims 6-8.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium is used to store program code, and the program code is used to execute the large-model-based geological prospecting literature knowledge extraction method described in any one of claims 1-5.

Images