Intergenerational artificial intelligence system based on multidimensional spatiotemporal information vector graphics

Abstract

The present invention relates to the field of artificial intelligence technology and technical applications and discloses an intergenerational artificial intelligence system based on multidimensional spatiotemporal information vector graphics. Based on a geographic information system (GIS) or a computer-aided design (CAD) platform and a data source, the present invention constructs a multidimensional spatiotemporal vector large-scale model endpoint, a multidimensional spatiotemporal information processing intelligence body endpoint, and an intelligent information system application endpoint, wherein, with respect to a two- or three-dimensional vector and the temporal and multidimensional spatiotemporal data, a multimodal spatiotemporal large-scale model is constructed with the capability to provide the system with technical expertise, data processing flow,The system is trained in advance to understand multidimensional vector graphics and topic-specific graphic documents, and the mutual expression and generation of technical multidimensional vector graphics and topic-specific graphic documents are realized to form the application of intelligent processing of technical graphic data. By establishing an artificial intelligence system in the field of engineering, the present invention will completely free engineers and technicians and offer intelligent system support for the automatic and rapid mutual processing of speech, text, and multidimensional vector graphics and topic-specific graphic documents in related technical fields.

Description

TECHNICAL AREA

[0001] The present invention relates to the field of artificial intelligence technology and technical applications, in particular an intergenerational artificial intelligence system based on multidimensional spatiotemporal information vector graphics. STATE OF THE ART

[0002] In recent years, artificial intelligence technology, represented by large-scale models, has developed rapidly. Based on training with massive datasets, large-scale modeling has demonstrated strong semantic understanding and inferential capabilities, showing potential for solving complex problems in fields such as language, translation, medicine, education, finance, and art. Multimodal large-scale models combine the natural language processing capabilities of large-scale models with the ability to understand and generate other modal data (e.g., visual, auditory data) to provide a richer and more natural interaction experience. This is achieved by integrating multiple types of input and output, such as text, images, and sound, to mimic the characteristics of multimodal information processing in human cognitive processes.Current AI technologies, including multimodal large-scale models, still deal with text, speech, image, video, and other data types, which are essentially text and raster formats, and are primarily used in literary and artistic creation, question-and-answer systems, language translation, and information summarization and analysis, while remaining largely unused in the application areas of human technology characterized by vector-format data. From birth, humans are confronted in the real world with rich multimedia spatiotemporal data, encompassing text, audio, video, graphics, images, and other types, as well as numbers, text, attributes, metadata, interaction processes, spatiotemporal relationships, and other complex features, which is why humans are intelligent and even wise.Language is only one organic part of multimedia spatiotemporal data, and the way humans characterize and interpret data consists primarily of text and vector and raster graphic formats. Artificial intelligence is not only the current textual or statistical intelligence or spatial intelligence based on machine vision, but also spatiotemporal intelligence (STAI). The large model will evolve into the spatiotemporal large model (STLM) in the future, and the large language model (LLM) is only a subset of the spatiotemporal large model (STLM). Vector models occupy an important position and are frequently used in spatiotemporal intelligent processing systems.

[0003] A Geographic Information System (GIS) is a computer system designed to capture, store, manage, calculate, analyze, display, and describe the spatial distribution of geographic features, as well as to solve user problems. Computer-aided design (CAD) is a computer system that uses computers and their graphics equipment to provide geometric modeling, feature calculation, computer mapping, and other functions for designers. GIS and CAD are widely used in spatial data management, engineering design, and cartography, and have accumulated a vast amount of information, including electronic maps, engineering drawings, production and construction drawings, etc., all of which are vector-based spatial data.Vector data is essentially an abstract data format that represents human spatial objects and technical processes. Whether in GIS or CAD, vector graphics are created through human interaction, and the resulting graphical documents are written by humans according to their knowledge base and the actual needs of the technology. Due to the presence of topological spatial relationships, GIS can achieve some degree of automated processing, whereas CAD lacks the description of spatiotemporal relationships and the complete expression of attributes, making intelligent processing difficult.Overall, the current mapping of GIS and CAD-related vector graphics and the processing of topic-specific graphic engineering documents are one-sided; this does not involve a bidirectional intergeneration of intelligence, therefore intelligent automatic processing cannot be realized.

[0004] Large-scale modeling offers the possibility of intergenerating vector graphics and graphic documents, but current research and application still focus primarily on text or raster data, including applications such as text-to-raster videos, while rarely considering the modeling of vector data such as points, lines, surfaces, solids, and networks. Therefore, there is an urgent need to develop a system for understanding, analyzing, and processing multidimensional spatiotemporal information vector graphics and for intergenerating topic-specific graphic documents, given the demand for processing large numbers of multidimensional vector graphics and subject-specific graphic documents in human-centered technical applications. CONTENT OF THE PRESENT INVENTION

[0005] With regard to the problems described above, the present invention develops an intergenerational artificial intelligence system based on multidimensional spatiotemporal information vector graphics.

[0006] One embodiment of the present invention provides an intergenerational artificial intelligence system based on multidimensional spatiotemporal information vector graphics, wherein the intergenerational artificial intelligence system comprises a multidimensional spatiotemporal vector large-scale model end, a multidimensional spatiotemporal information processing intelligence body end, and an intelligent information system application end;

[0007] and wherein the multidimensional spatiotemporal vector large-scale model end is used for a unified modeling expression of multiple types of spatiotemporal data to construct a multimodal spatiotemporal vector large-scale model, and wherein the spatiotemporal vector large-scale model is used for mutual expression, understanding, and analysis of multidimensional spatiotemporal vector data and natural language descriptions, including speech and text, to achieve enhanced learning of knowledge of vertical domains of the system for technical expertise and thus enhance the ability to process technical expertise and data processing, and wherein the multiple types of spatiotemporal data include text, speech, image, video, and multidimensional vector graphics;

[0008] and wherein the multidimensional spatiotemporal information processing intelligence body end is used to learn and consult the task execution process of a typical business process based on the learning and inference capability of the spatiotemporal vector grand model and in combination with the processing and analysis capability of a geographic information system (GIS) or a computer-aided design (CAD) platform, to automatically analyze and divide the business process tasks of various types of multidimensional spatiotemporal data processing interactions into simple subtasks, and to autonomously execute, transfer, and solve the subtasks;

[0009] and wherein the intelligent information system application end is used to convert various types of multidimensional spatiotemporal data into corresponding natural language descriptions based on the inference and understanding capability of the large-scale spatiotemporal vector model, in order to realize the understanding of multidimensional vector graphics and various types of multidimensional spatiotemporal data by artificial intelligence, to complete the interaction, analysis and intergeneration of intelligent body-controlled multidimensional vector graphics and topic-specific graphic documents in the field of technology by means of simple inputs including speech and text,and through the inferential capability of the large-scale spatial-temporal vector model and the processing capability of the intelligence body, to accurately capture the parameters in the multidimensional spatial-temporal data and to automatically update the relevant multidimensional spatial-temporal data, thereby realizing the automatic generation and updating of the multidimensional vector graphics or the topic-specific graphic documents.

[0010] Optionally, the system's domain knowledge for technical expertise is embedded in the multidimensional spatiotemporal vector large-scale model end to achieve enhanced learning of vertical domains of the system for technical expertise and to strengthen the ability to understand and perform technical domain knowledge and data processing processes;

[0011] wherein the process in which the multidimensional spatiotemporal vector large-scale model end constructs the spatiotemporal vector large-scale model includes: natural language description of the multidimensional spatiotemporal vector data, pre-training and fine-tuning of the spatiotemporal vector large-scale model, and understanding and analyzing the multidimensional spatiotemporal vector data;

[0012] and wherein the fact that the domain knowledge of the technical expertise system is embedded into the multidimensional spatiotemporal vector grand model end to achieve enhanced learning of the knowledge of vertical domains of the technical expertise system includes: fine-tuning the generic grand model with technical domain data as a pre-training dataset, and pre-training the generic grand model by extending the vertical technical domain lexicon using the technical domain dataset to embed the domain knowledge of the technical expertise system into the generic grand model;

[0013] and wherein the fact that the domain knowledge of the graphical technical data processing processes is embedded into the multidimensional spatiotemporal vector large-scale model end, in order to enhance the ability to understand and execute technical domain knowledge and data processing processes, includes in particular the following: describing the data processing processes with clear technical features, defining the language of the technical domain data processing processes, training the generic large-scale model with the technical domain data processing processes, which are expressed uniformly, as a fine-tuning dataset, in order to realize the knowledge embedding of the data processing processes and to enhance the ability of the generic large-scale model to understand and execute the technical domain data processing processes.

[0014] and wherein the natural language description of the multidimensional spatiotemporal vector data includes, but is not limited to, spatiotemporal vector data of various types of points, lines, surfaces and solids with continuous spatiotemporal (x, y, z, t) information and attribute information in various domains, including geographic information systems (GIS) and computer-aided mapping (CAD), and wherein, according to the requirements of training, understanding and processing the large-scale spatiotemporal vector model, the description transformation is performed in the form of natural language, and wherein the transformed natural language description contains all the information of the original spatiotemporal vector data, including, but not limited to, the geometric type of the vector object, the geometric features, the display style, the coordinates of the reference point,the relative geometric data based on the reference point, the temporal information, and the attribute information;

[0015] and wherein the pre-training and fine-tuning of the spatiotemporal vector large-scale model includes: creating a dataset that includes the natural language descriptions of the multidimensional spatiotemporal vector data and the vector object features, continuous pre-training and model fine-tuning on the generic large-scale model based on the pre-training dataset and the fine-tuning dataset, and forming a spatiotemporal vector large-scale model with the ability to understand the spatiotemporal vector data;

[0016] and wherein understanding and analyzing the multidimensional spatiotemporal vector data includes: inputting spatiotemporal vector data described in natural language into the spatiotemporal vector macromodel, wherein the spatiotemporal vector macromodel outputs features of the understood vector object, and inputs some of the features of the vector object into the spatiotemporal vector macromodel, and wherein the spatiotemporal vector macromodel outputs complete spatiotemporal vector data that is understood or analyzed, and based on the output spatiotemporal features or data of the vector object, understanding and processing geometric features of the spatiotemporal vector data, understanding and processing attribute features,the differentiation and processing of spatial relationships and the differentiation and processing of spatiotemporal relationships are further complemented by the analytical capabilities of the spatiotemporal vector macromodel;

[0017] and wherein understanding and processing geometric features of the spatiotemporal vector data includes: processing geometric coordinates of the vector object itself, including but not limited to modifying, editing and manipulating shapes, sizes and positions;

[0018] and wherein understanding and processing attribute features of the spatiotemporal vector data includes: processing attribute information of the vector object itself, including but not limited to processing modification, queries, analysis and statistics;

[0019] and wherein the distinguishing and processing of spatial relationships of the spatiotemporal vector data includes: distinguishing and processing the topological spatial relationship, the sequential spatial relationship, and the metric spatial relationship between the vector objects; wherein the topological spatial relationship refers to the association, proximity, inclusion, intersection, overlap, and separation relationship between the spatial objects; wherein the sequential spatial relationship refers to the spatial ordering of the spatial objects or events in space, including front-back, left-right, top-bottom, and east-west-north-south orientation relationships; and wherein the metric spatial relationship refers to a distance or proximity relationship between spatial objects.

[0020] Optionally, the multidimensional spatiotemporal information processing intelligence body end includes a task planning module, a task memory module, and a task action module; wherein the task planning module, the task memory module, and the task action module each represent a corresponding processor.

[0021] The task planning module is used to decompose and schedule the subtasks of interactive vector graphics business processes, contained in various types of multidimensional spatiotemporal data, based on domain knowledge, business processes, and the capabilities of the large-scale spatiotemporal vector model for understanding data and generating text. The respective subtasks are general logic processing or spatial analysis operations on geographic information, executed independently with clear inputs and outputs. The respective subtasks are then executed in a linked manner to complete a complex business process.

[0022] and wherein the task memory module is used such that each task gathers information from the environment or receives information from memory to provide the necessary data for each task and to store the process data during execution; and wherein, through the task memory module, the multidimensional spatiotemporal information vector processing intelligence body gathers the data and experience and gradually completes self-evolution to provide support for the iterative capability of the spatiotemporal vector grand model;

[0023] and wherein the task action module is used to execute each subtask planned by the spatiotemporal vector large-scale model into a specific result; and wherein the task execution process relies on, but is not limited to, the inferential capability of the spatiotemporal vector large-scale model, the spatial analysis and processing capability of the geographic information system, and the processing capability of the computer-aided mapping system, and wherein the objects interacting in the execution process include, but are not limited to, sensors, model libraries, controllers, and databases.

[0024] Optionally, the step in which artificial intelligence realizes the understanding of multidimensional vector graphics and various types of multidimensional spatiotemporal data, and performs interaction, analysis, and intergeneration on the multidimensional vector graphics and the topic-specific graphic documents, includes the following: processing and understanding multidimensional spatiotemporal data, generating a multidimensional spatiotemporal data processing intelligence body, and intergenerating multidimensional vector graphics and topic-specific graphic documents, which in detail includes the following steps:

[0025] Step S1: Inputting the multidimensional spatiotemporal data into the intelligent information system application, converting the multidimensional spatiotemporal data into natural language descriptions that can be processed by the spatiotemporal vector model, realizing the analysis and understanding of multidimensional spatiotemporal information vector data and temporal data within the multidimensional spatiotemporal data by artificial intelligence based on the reasoning and understanding capabilities of the spatiotemporal vector model;

[0026] Step S2: Input of user instructions through interactive means, including speech input and text input, with speech input being converted into text by speech recognition; conversion of the user instructions described in natural language into formatted information system data based on the understanding, analysis, and processing capabilities of the spatiotemporal vector grand model; editing, querying, analyzing, and outputting the instructions to realize the generation of the multidimensional spatiotemporal data processing intelligence body;

[0027] Step S3: based on the multidimensional spatiotemporal information processing intelligence body, using its fusion call and multiple execution rounds, realizing the automatic generation of automated and intelligent multidimensional vector graphics and theme-specific graphic documents and the bidirectional updating of intergenerational artificial intelligence applications, and realizing the intergeneration of multidimensional vector graphics and various types of multidimensional spatiotemporal data.

[0028] Optionally, the creation of a dataset containing the natural language descriptions of the multidimensional spatiotemporal vector data and the vector object features, and continuous pretraining and model fine-tuning on the generic large-scale model based on the pretraining dataset and the fine-tuning dataset, include in particular the following steps:

[0029] Step T1: Collecting a huge amount of spatiotemporal vector data, the corresponding natural language text descriptions of spatiotemporal vector data and the feature descriptions of the vector object, and forming a spatiotemporal vector training dataset, where the spatiotemporal vector data includes predefined geometric structure information of points, lines, surfaces and bodies, as well as attribute text information corresponding to the geometric bodies;

[0030] Step T2: Performing pre-training with the generic large-scale model as a basis using the spatiotemporal vector training dataset, performing vector text sampling by random wandering as a training sample, and fine-tuning the model with generative GPT input into the generic large-scale model to achieve effective injection of vector information knowledge;

[0031] Step T3: Collect several types of question and answer pattern data, including vector data topological relationship, vector data attribute information, attribute description vector data, to form a spatiotemporal vector model fine-tuning dataset, and perform fine-tuning of the pre-trained generic large-scale model in step T2 to achieve effective alignment between natural language and vector data.

[0032] Optionally, the task scheduling module retrieves similar tasks from a technical process library based on a task description, wherein the task scheduling mode in the task scheduling module includes scheduling without feedback and scheduling with feedback, and wherein the feedback comes from the environment, the user, or the execution result of the spatiotemporal vector macromodel, and wherein the business process is decomposed into a multitude of subtasks to be executed, and wherein the multitude of subtasks are linked together in a cascade or tree-like manner, and wherein the subsequent subtasks are determined after completion of each subtask based on the results of the task execution.

[0033] Optionally, the task memory module stores information perceived from the environment, a record of task execution, the result of task execution, and uses the recorded memories to promote future actions, with the task memory module encompassing stored data and operations;

[0034] and wherein the stored data comprises a short-term memory of input information within a context window and a long-term memory of external vector memory, which is retrieved by a quick query.

[0035] Optionally, the task action module uses tools to complete the execution of individual tasks, wherein the specific tools include the large-scale spatiotemporal vector model itself and external tools, including algorithmic models, program compilations, databases, and APIs, and wherein the task action module further uses spatial databases to perform some of the spatial analyses, uses models in the model library to complete specific tasks, and uses APIs to obtain real-time or historical data, and wherein the specific tasks are tasks corresponding to individual models in the model library.

[0036] Optionally, the implementation of the automatic generation of multidimensional vector graphics or theme-specific graphic documents includes the following:

[0037] Artificial formulation of a template for the multidimensional vector graphics or the topic-specific graphic documents based on the content requirements of the multidimensional vector graphics and the topic-specific graphic documents, or automatic generation of a topic-specific graphic document template by the spatial-temporal vector macromodel based on the content requirements, wherein the graphics in the topic-specific graphic documents are in a multidimensional vector graphics format or in a raster format that has been converted from a multidimensional vector graphic;

[0038] Extracting and generating various types of parameters in the template of multidimensional vector graphics or topic-specific graphic documents by querying and analyzing the information system or understanding and inferring the spatiotemporal vector macro-model, accurately capturing parameters in the multidimensional spatiotemporal vector data, merging the parameters with the template and generating the multidimensional vector graphics or topic-specific graphic documents.

[0039] Optionally, the intergeneration of the multidimensional vector graphics and the topic-specific graphic documents is based on the understanding and analysis capabilities of the spatiotemporal vector model, whereby, after the relevant changes to the spatiotemporal information caused by the technical domain data or the topic-specific graphic documents have been entered, the spatiotemporal vector model generates the updated parameters, descriptions, or intelligence bodies of the multidimensional vector graphics in order to automatically update the data of the multidimensional vector graphics;

[0040] and wherein, after the mapping or modification of the multidimensional vector graphics has led to a change in the multidimensional vector graphics, the modified multidimensional vector graphics are captured by the spatial-temporal vector macromodel and the corresponding technical parameters or document descriptions are updated in order to update the relevant content of the topic-specific graphic documents and to realize the intergeneration of the multidimensional vector graphics and the topic-specific graphic documents.

[0041] Based on a geographic information system (GIS) or a computer-aided design (CAD) platform and a data source, the present invention constructs a multidimensional spatiotemporal vector large-scale model end, a multidimensional spatiotemporal information processing intelligence body end, and an intelligent information system application end, wherein, with respect to a two- or three-dimensional vector and the temporal and multidimensional spatiotemporal data, a large-scale model is pre-trained with the ability to understand the system for technical expertise, the data processing flow, the multidimensional vector graphics, and the subject-specific document structures, and wherein the mutual expression and generation of the multidimensional vector graphics, the subject-specific spatiotemporal information, and the professional document description are realized.to form the application of the intelligent technical domain data processing system. By establishing an artificial intelligence system in the field of human engineering, the present invention will completely free engineers and technicians and offer intelligent system support for the automatic and rapid mutual processing of speech, text, and multidimensional vector graphics and topic-specific graphic documents in related fields, resulting in improved practicality. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] Several other advantages and benefits will become apparent to the person skilled in the art upon reading the following detailed description of the preferred embodiments. The accompanying drawings serve only to illustrate the preferred embodiments and are not considered to limit the present invention. Furthermore, the same components are designated by the same reference numerals in all drawings. In the drawings, the following are shown: Fig. 1 a structural block diagram of an intergenerational artificial intelligence system based on multidimensional spatiotemporal information vector graphics, provided by an embodiment of the present invention; Fig. 2 a flowchart of an application of an intelligent information system application end in an embodiment of the present invention based on the spatiotemporal vector macromodel; Fig. 3 a flowchart of the realization of interacting, analyzing and intergenerating vector graphics and graphic documents in an embodiment of the present invention based on the spatiotemporal vector macromodel; Fig. 4 a schematic diagram of another embodiment of the intergenerational artificial intelligence system based on multidimensional spatiotemporal information vector graphics, provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0043] The present invention is explained in more detail below in conjunction with figures and detailed embodiments, so that the stated objective, features, and advantages of the present invention are clearer and easier to understand. It is understood that the detailed embodiments described here serve only to illustrate the present invention and represent only a part of the embodiments of the present invention, rather than all embodiments, and are not intended to limit the present invention.

[0044] See Fig. Figure 1 shows a structural block diagram of an intergenerational artificial intelligence system based on multidimensional spatiotemporal information vector graphics in the present embodiment, wherein the intergenerational artificial intelligence system comprises, in particular, a multidimensional spatiotemporal vector large-scale model end, a multidimensional spatiotemporal information processing intelligence body end, and an intelligent information system application end. The intelligence body (agent) is an independent small module, a small application capable of performing a separate task; and an information processing intelligence body is a small module for information processing.

[0045] The multidimensional spatiotemporal vector large-scale model end is used for a unified modeling representation of several types of spatiotemporal data, i.e., for a unified modeling representation of different types of spatiotemporal data. These types of spatiotemporal data include: text, speech, images, video, and multidimensional vector graphics.

[0046] The multidimensional spatiotemporal vector large-scale model end is further used to construct a multimodal spatiotemporal vector large-scale model, whereby the multimodal spatiotemporal vector large-scale model is used for mutual expression, understanding and analysis of multidimensional spatiotemporal vector data and descriptions in natural language, including speech and text, in order to achieve enhanced learning of knowledge of vertical domains of the system for technical expertise and thus enhance the ability to process technical expertise and data processing.

[0047] Preferably, the system's domain knowledge for technical expertise is embedded in the multidimensional spatiotemporal vector large-scale model end to achieve enhanced learning of knowledge from vertical domains of the system for technical expertise and to strengthen the ability to understand and perform technical domain knowledge and data processing processes;

[0048] The process by which the multidimensional spatiotemporal vector large-scale model is constructed includes the following: natural language description of the multidimensional spatiotemporal vector data, pre-training and fine-tuning of the spatiotemporal vector large-scale model, and understanding and analyzing the multidimensional spatiotemporal vector data. Fig. 1. The elements with which the multidimensional spatiotemporal vector large-scale model end constructs the spatiotemporal vector large-scale model are shown as examples to better illustrate the structure of the intergenerational system of artificial intelligence, which does not mean that the multidimensional spatiotemporal vector large-scale model end is only used to construct the spatiotemporal vector large-scale model.The elements with which the multidimensional spatiotemporal vector large-scale model end constructs the spatiotemporal vector large-scale model include: natural language descriptions of the spatiotemporal vector data (namely natural language descriptions of the multidimensional spatiotemporal vector data), pretraining and fine-tuning of the spatiotemporal vector large-scale model (namely pretraining and fine-tuning of the multidimensional spatiotemporal vector large-scale model), and understanding and analyzing the spatiotemporal vector data (namely understanding and analyzing the multidimensional spatiotemporal vector data).

[0049] The fact that the domain knowledge of the technical expertise system is embedded into the multidimensional spatiotemporal vector large-scale model end to achieve enhanced learning of the system's vertical domains includes, in particular, the following: fine-tuning of the generic large-scale model with technical domain data as a pre-training dataset, and pre-training of the generic large-scale model by extending the vertical technical domain lexicon using the technical domain dataset to embed the domain knowledge of the technical expertise system into the generic large-scale model;The generic grand model, also called the "basic grand model," refers to the large language model, which can be domain- and task-spanning, is generally provided through the training of large organizations, and pays more attention to the general capability and generalizability of the model; accordingly, the spatiotemporal grand model in the description is the domain- and vertical grand model, in which the direction-specific capabilities are extended on the basis of the generic grand model.

[0050] The fact that the domain knowledge of graphical technical data processing processes is embedded into the multidimensional spatiotemporal vector large-scale model end, in order to enhance the ability to understand and execute technical domain knowledge and data processing processes, includes in particular the following: describing the data processing processes with clear technical characteristics, defining the language of the technical domain data processing processes, training the generic large-scale model with the technical domain data processing processes, which are expressed uniformly, as a fine-tuning dataset, in order to realize the knowledge embedding of the data processing processes and to enhance the ability of the generic large-scale model to understand and execute the technical domain data processing processes.

[0051] The natural language description of multidimensional spatiotemporal vector data includes, but is not limited to, spatiotemporal vector data of various types of points, lines, surfaces, and solids with continuous spatiotemporal (x, y, z, t) information and attribute information in various fields, including geographic information systems (GIS) and computer-aided mapping (CAD). The description is transformed into natural language according to the requirements of training, understanding, and processing the large-scale spatiotemporal vector model. The transformed natural language description contains all the information of the original spatiotemporal vector data, including, but not limited to, the geometric type of the vector object, its geometric features, the display style, and the coordinates of the reference point.the relative geometric data based on the reference point, the temporal information, and the attribute information.

[0052] The pre-training and fine-tuning of the spatiotemporal vector large-scale model include, in particular, the following: creating a dataset that includes the natural language descriptions of the multidimensional spatiotemporal vector data and the vector object features, continuous pre-training and model fine-tuning on the generic large-scale model based on the pre-training dataset and the fine-tuning dataset, and forming a spatiotemporal vector large-scale model with the ability to understand the spatiotemporal vector data.

[0053] Preferably, the creation of a dataset that includes the natural language descriptions of the multidimensional spatiotemporal vector data and the vector object features, and continuous pretraining and model fine-tuning on the generic large-scale model based on the pretraining dataset and the fine-tuning dataset, preferably include the following steps:

[0054] Step T1: Collecting a huge amount of spatiotemporal vector data, the corresponding natural language text descriptions of spatiotemporal vector data and the feature descriptions of the vector object, and forming a spatiotemporal vector training dataset, where the spatiotemporal vector data includes predefined geometric structure information of points, lines, surfaces and bodies, as well as attribute text information corresponding to the geometric bodies;

[0055] Step T2: Performing pre-training with the generic large-scale model as a basis using the spatiotemporal vector training dataset, performing vector text sampling by random wandering as a training sample, and fine-tuning the model with generative GPT input into the generic large-scale model to achieve effective injection of vector information knowledge;

[0056] Step T3: Collect several types of question and answer pattern data, including vector data topological relationship, vector data attribute information, attribute description vector data, to form a spatiotemporal vector model fine-tuning dataset, and perform fine-tuning of the pre-trained generic large-scale model in step T2 to achieve effective alignment between natural language and vector data.

[0057] The procedure with the three steps above ensures that the multidimensional spatiotemporal vector macromodel end constructs the spatiotemporal vector macromodel.

[0058] Understanding and analyzing multidimensional spatiotemporal vector data includes the following: inputting spatiotemporal vector data described in natural language into the spatiotemporal vector model, where the spatiotemporal vector model outputs features of the understood vector object, and inputs some of the features of the vector object into the spatiotemporal vector model, and where the spatiotemporal vector model outputs complete spatiotemporal vector data that has been understood or analyzed, and, based on the output spatiotemporal features or data of the vector object, understanding and processing geometric features of the spatiotemporal vector data, understanding and processing attribute features,The differentiation and processing of spatial relationships and the differentiation and processing of spatiotemporal relationships are further complemented by the analytical capabilities of the spatiotemporal vector macromodel.

[0059] Understanding and processing geometric features of spatiotemporal vector data includes: processing geometric coordinates of the vector object itself, including but not limited to modifying, editing, and manipulating shapes, sizes, and positions; understanding and processing attribute features of spatiotemporal vector data includes: processing attribute information of the vector object itself, including but not limited to processing modification, queries, analysis, and statistics; distinguishing and processing spatial relationships of spatiotemporal vector data includes: distinguishing and processing the topological spatial relationship, the sequential spatial relationship, and the metric spatial relationship between the vector objects;where the topological spatial relationship refers to the association, proximity, inclusion, intersection, overlap, and separation relationship between spatial objects; the sequential spatial relationship refers to the spatial arrangement order of spatial objects or events in space, including front-back, left-right, top-bottom, and east-west-north-south orientation relationships; and the metric spatial relationship refers to a distance or proximity relationship between spatial objects.

[0060] The multidimensional spatiotemporal information processing intelligence body is used to learn and consult the task execution process of a typical business process, based on the learning and inference capabilities of the spatiotemporal vector macro-model and in combination with the processing and analysis capabilities of a geographic information system (GIS) or a computer-aided design (CAD) platform. It automatically analyzes the business process tasks of various types of multidimensional spatiotemporal data processing interactions, divides them into simple subtasks, and autonomously executes, transfers, and solves these subtasks. The typical business process refers to the process and steps for processing the aforementioned spatiotemporal information data and encompasses GIS- and CAD-related data processing methods, such as...Mapping graphs, editing, adding attributes, processing spatial analyses, etc.

[0061] Preferably, the multidimensional spatiotemporal information processing intelligence body end comprises a task planning module, a task memory module, and a task action module.

[0062] The task planning module is used to decompose and plan the subtasks of interactive vector graphics business processes, contained in various types of multidimensional spatiotemporal data, based on domain knowledge, business processes, and the capabilities of the large-scale spatiotemporal vector model to understand the data and generate the text. The respective subtasks are general logic processing or spatial analysis operations of geographic information, executed independently with clear inputs and outputs. The respective subtasks can be executed in a linked manner to complete complex business process processing.In particular, the task scheduling module can first retrieve similar tasks from a technical process library based on a task description, wherein the task scheduling mode in the task scheduling module includes scheduling without feedback and scheduling with feedback, and wherein the feedback may come from the environment, the user, or the execution result of the spatiotemporal vector macromodel, and wherein the business process is decomposed into a multitude of subtasks to be executed, and wherein the multitude of subtasks are linked together in a cascade or tree-like manner, and wherein the subsequent subtasks are determined after completion of each subtask based on the results of the task execution.

[0063] The task memory module is used to allow individual tasks to gather information from the environment or retrieve information from memory to provide data required for individual tasks and to store process data during execution. With the help of the task memory module, the multidimensional spatiotemporal information vector processing intelligence body can gather data and experience and gradually undergo self-evolving, thus providing iterative support for the spatiotemporal vector grand model.In particular, the task memory module can store information perceived from the environment, a record of task execution, the result of task execution, and use the recorded memories to promote future actions, wherein the task memory module comprises stored data and operations; the stored data includes a short-term memory of input information within a context window and a long-term memory of external vector memory, which is retrieved by a quick query.

[0064] The task action module is used to execute each subtask planned by the spatiotemporal vector model into a specific result. The task execution process relies on, but is not limited to, the inferential capabilities of the spatiotemporal vector model, the spatial analysis and processing capabilities of the geographic information system, and the processing capabilities of the computational mapping system. The objects interacting in the execution process include, but are not limited to, sensors, model libraries, controllers, and databases. In particular, the task action module can use tools to complete the execution of individual tasks. These specific tools include the spatiotemporal vector model itself and external tools, including algorithmic models, program compilations, databases, APIs, and so on.include, and wherein the task action module can further use spatial databases to perform some of the spatial analyses, use models in the model library to complete specific tasks, and use APIs to obtain real-time or historical data, etc., and wherein the specific tasks are tasks that correspond to individual models in the model library.

[0065] In Fig. 1. Three modules of the multidimensional spatiotemporal information processing intelligence body end are presented as examples to better illustrate the structure of the intergenerational artificial intelligence system. This does not mean that the multidimensional spatiotemporal information processing intelligence body end comprises only these three modules. The three modules of the multidimensional spatiotemporal information processing intelligence include: task planning (namely a task planning module), task memory (namely a task memory module), and task action (namely a task action module).

[0066] The intelligent information system application is used to transform various types of multidimensional spatiotemporal data into corresponding natural language descriptions based on the inference and understanding capabilities of the large-scale spatiotemporal vector model. This enables artificial intelligence to understand multidimensional vector graphics and various types of multidimensional spatiotemporal data, and to complete the interaction, analysis, and intergeneration of intelligent, body-controlled multidimensional vector graphics and topic-specific graphic documents in the field of technology using simple inputs including speech and text.and through the inferential capability of the spatiotemporal vector macromodel and the processing capability of the intelligence body, the parameters in the multidimensional spatiotemporal data are accurately captured, and the relevant multidimensional spatiotemporal data are automatically updated, thereby realizing the automatic generation and updating of the multidimensional vector graphics or the topic-specific graphic documents. In the exemplary embodiments of the present invention, a multidimensional vector graphic refers to two-dimensional, three-dimensional, and time-dimensional vector data used in a technical field; and a topic-specific graphic document refers to a document report with a mixture of graphics and text with a technically specific meaning or application.

[0067] Preferably, the step in which artificial intelligence realizes the understanding of multidimensional vector graphics and various types of multidimensional spatiotemporal data, and performs interaction, analysis, and intergeneration on the multidimensional vector graphics and the topic-specific graphic documents, comprises the following: processing and understanding multidimensional spatiotemporal data, generating a multidimensional spatiotemporal data processing intelligence body, and intergenerating multidimensional vector graphics and topic-specific graphic documents, where the multidimensional spatiotemporal data processing intelligence body refers to an intelligence body that processes the multidimensional spatiotemporal data, such as reading, editing, analyzing, etc., and specifically includes the following steps:

[0068] Step S1: Inputting the multidimensional spatiotemporal data into the intelligent information system application, converting the multidimensional spatiotemporal data into natural language descriptions that can be processed by the spatiotemporal vector model, realizing the analysis and understanding of multidimensional spatiotemporal information vector data and temporal data within the multidimensional spatiotemporal data by artificial intelligence based on the reasoning and understanding capabilities of the spatiotemporal vector model;

[0069] Step S2: Input of user instructions through interactive means, including speech input and text input, with speech input being converted into text by speech recognition; conversion of the user instructions described in natural language into formatted information system data based on the understanding, analysis, and processing capabilities of the spatiotemporal vector grand model; editing, querying, analyzing, and outputting the instructions to realize the generation of the multidimensional spatiotemporal data processing intelligence body;

[0070] Step S3: based on the multidimensional spatiotemporal information processing intelligence body, using its fusion call and multiple execution rounds, realizing the automatic generation of automated and intelligent multidimensional vector graphics and theme-specific graphic documents and the bidirectional updating of the applications of intergenerational artificial intelligence, and realizing the intergeneration of multidimensional vector graphics and theme-specific graphic documents.

[0071] In Fig. 1. The elements with which the intelligent information system application end realizes the understanding of multidimensional vector graphics and various types of multidimensional spatiotemporal data by artificial intelligence and performs the interaction, analysis and intergeneration on the multidimensional vector graphics and the topic-specific graphic documents are presented as examples in order to better illustrate the structure of the intergenerational system of artificial intelligence, which does not mean that the intelligent information system application end only completes these technical contents.The elements include: processing and understanding multidimensional spatiotemporal data (namely, analyzing and understanding multidimensional spatiotemporal information vector data and temporal data within the multidimensional spatiotemporal data by artificial intelligence), generating a multidimensional spatiotemporal data processing intelligence body (namely, generating a multidimensional spatiotemporal data processing intelligence body), and intergenerating multidimensional spatiotemporal information vector graphics and topic-specific graphic documents (namely, intergenerating multidimensional vector graphics and topic-specific graphic documents).

[0072] Implementing the automatic generation of multidimensional vector graphics or theme-specific graphic documents includes, in particular, the following:

[0073] Firstly, the artificial formulation of a template for the multidimensional vector graphics or the topic-specific graphic documents based on the content requirements of the multidimensional vector graphics and the topic-specific graphic documents, or automatic generation of a topic-specific graphic document template by the spatial-temporal vector large-scale model based on the content requirements, whereby the graphics in the topic-specific graphic documents can be in both a multidimensional vector graphics format and a raster format that has been converted from a multidimensional vector graphic.Subsequent extraction and generation of various types of parameters in the template of the multidimensional vector graphics or the topic-specific graphic documents by querying and analyzing the information system or understanding and inferring the spatiotemporal vector macro-model, accurate capture of parameters in the multidimensional spatiotemporal vector data, and finally merging the parameters with the template and generating the multidimensional vector graphics or the topic-specific graphic documents.

[0074] The intergeneration of the multidimensional vector graphics and the topic-specific graphic documents is based on the understanding and analysis capabilities of the spatiotemporal vector model, whereby, after the relevant changes to the spatiotemporal information caused by the technical domain data or the topic-specific graphic documents have been entered, the spatiotemporal vector model generates the updated parameters, descriptions, or intelligence bodies of the multidimensional vector graphics in order to automatically update the data of the multidimensional vector graphics;

[0075] and wherein, after the mapping or modification of the multidimensional vector graphics has led to a change in the multidimensional vector graphics, the modified multidimensional vector graphics are captured by the spatial-temporal vector macromodel and the corresponding technical parameters or document descriptions are updated in order to update the relevant content of the topic-specific graphic documents and to realize the intergeneration of the multidimensional vector graphics and the topic-specific graphic documents.

[0076] Fig. Figure 2 shows a flowchart of an application of an intelligent information system application in the present embodiment based on the spatiotemporal vector macromodel, wherein, based on the spatiotemporal vector macromodel, the mutual expression and generation of natural language, such as speech input, text input, etc., and of multidimensional vector graphics and topic-specific graphic documents, etc., are realized. The "understanding and analysis of the vector image and the temporal state," the "instruction for applying the vector image and the temporal state," and the "information processing intelligence body" in Fig. 2 can be understood according to the steps S1 to S3 mentioned above and are not repeated here.

[0077] Fig. Figure 3 shows a flowchart illustrating the implementation of interacting, analyzing, and intergenerating multidimensional vector graphics and topic-specific graphic documents in the present embodiment, based on the spatiotemporal vector model. Based on the understanding and analysis capabilities of the spatiotemporal vector model, various types of multidimensional spatiotemporal data, such as technical domain data, graphic documents, and other spatiotemporal data sources, are intergenerated and updated with different types of vector graphics (2D graphics, 3D graphics, and other vector types, etc.) and data (vector graphic line parameters, vector temporal parameters, vector graphic descriptions, and vector graphic intelligence). The vector graphic intelligence refers specifically to a small module that handles the tasks of the vector graphics (e.g., reading, editing, saving, etc.).

[0078] In summary, the present invention constructs, based on a geographic information system (GIS) or a computer-aided design (CAD) platform and a data source, a multidimensional spatiotemporal vector large-scale model end, a multidimensional spatiotemporal information processing intelligence body end, and an intelligent information system application end, wherein, with respect to a two- or three-dimensional vector and the temporal and multidimensional spatiotemporal data, a large-scale model is pre-trained with the ability to understand the system for technical expertise, the data processing flow, the multidimensional vector graphics, and the subject-specific document structures, and wherein the mutual expression and generation of the multidimensional vector graphics, the subject-specific spatiotemporal information, and the professional document description are realized.to form the application of the intelligent technical domain data processing system. By establishing an artificial intelligence system in the field of human engineering, the present invention will completely free engineers and technicians and offer intelligent system support for the automatic and rapid mutual processing of speech, text, and multidimensional vector graphics and topic-specific graphic documents in related fields, resulting in improved practicality.

[0079] The intergenerational artificial intelligence system described above, based on multidimensional spatiotemporal information vector graphics, is merely an example. A unit described as a separate element may be physically separate or not, and an element displayed as a unit may be a physical unit or not; that is, it may be located in one place or distributed across multiple network units. Depending on actual needs, some or all of the modules can be selected to achieve the goal of the solution in this embodiment. The average person skilled in the art can understand and implement the present invention without any creative effort.

[0080] The various components of this disclosure can be implemented in hardware or in software modules running on one or more processors, or in combinations thereof. Those skilled in the art should understand that microprocessors or digital signal processors (DSPs) can be used in practice to implement some or all of the functions of some or all of the components of an electronic device according to the embodiments of this disclosure. This disclosure can also be implemented in the form of a device or a program of equipment (e.g., computer programs and computer program products) for carrying out some or all of the methods described herein. Such a program implementing this disclosure can be stored on a computer-readable medium or may be in the form of one or more signals.Such signals may be available for download from a website or provided on a carrier signal or in any other form.

[0081] For example, it shows Fig. 4. An intergenerational artificial intelligence system based on multidimensional spatiotemporal information vector graphics that can implement the present disclosure. The intergenerational artificial intelligence system based on multidimensional spatiotemporal information vector graphics traditionally comprises a processor 101 and a computer program product or a computer-readable medium in the form of a memory 1020. The memory 1020 can be an electronic memory such as flash memory, EEPROM (electrically erasable programmable read-only memory), EPROM, a hard disk, or ROM. The memory 1020 has a memory location 1030 for program code 1031 for carrying out one of the process steps in the procedure described above. For example, the memory location 1030 for program code can contain individual program codes 1031 for implementing the various steps in the procedure described above.These program codes can be read from or written to one or more computer program products. These computer program products include program code carriers such as hard drives, compact discs (CDs), memory cards, or floppy disks. Such computer program products are typically portable or fixed storage units. The storage unit may have memory segments, storage space, etc., similar to the memory 1020 in the intergenerational system of artificial intelligence based on multidimensional spatiotemporal information vector graphics of [missing information]. Fig.The program code can be compressed in a suitable form. Typically, the memory unit includes computer-readable code, i.e., code that can be read by a processor such as a 1010 and which, when executed by an electronic device, causes the electronic device to perform the various steps of the procedure described above.

[0082] It should be noted that for the relevant content in the present embodiment, reference can be made to the preceding embodiments concerning the intergenerational system of artificial intelligence based on multidimensional spatiotemporal information vector graphics, which is not repeated here.

[0083] Although the preferred embodiments of the present invention have already been explained in detail, a person skilled in the art in this field can make other modifications and changes to these embodiments once they are familiar with the essential creative concepts. Therefore, the claims should be understood to encompass the preferred embodiments and all changes and modifications falling within the scope of the embodiments of the present invention.

[0084] Finally, it should be noted that relational terms such as "first," "second," etc., in the description serve only to distinguish one object or activity from another, without necessarily requiring or implying that any such actual relationship or sequence exists between the objects or activities. Furthermore, terms such as "include," "exhibit," or other variations cover non-exclusive exhibiting, so that a process, procedure, object, or terminal encompassing a series of elements includes both such elements and other elements not explicitly listed or inherent to that process, procedure, object, or terminal. If no further restrictions apply, a process or procedure described as "encompassing a series of elements" will...The defined element does not exclude the possibility that other identical elements exist within a process, procedure, object or device that encompasses the element.

[0085] The embodiments of the present invention are explained in more detail above in connection with the figures; however, the present invention is not limited to the embodiments described above. The specific embodiments mentioned above are merely illustrative and not limiting. Inspired by the present invention, the person skilled in the art in this field can, without deviating from the principle of the scope of protection of the claims, also implement many forms, all of which fall within the scope of protection of the present invention.

Claims

[1] Intergenerational artificial intelligence system based on multidimensional spatiotemporal information vector graphics, comprising a processor and a memory, wherein instructions are stored in the memory, and wherein the processor executes the instructions, and wherein the processor is configured to do so: to express several types of spatiotemporal data uniformly and to construct a multimodal spatiotemporal vector grand model, wherein the spatiotemporal vector grand model is used to understand and analyze the natural language descriptions of the multidimensional spatiotemporal vector data in order to enhance the ability to perform technical domain knowledge and data processing processes, and wherein the types of spatiotemporal data include: text, speech, image, video and multidimensional vector graphics, and wherein the natural language descriptions of multidimensional spatiotemporal vector data include speech and text; based on the learning and reasoning capabilities of the large-scale spatial-temporal vector model and in combination with the processing and analysis capabilities of a geographic information system (GIS) or a computer-aided design (CAD) platform, to learn and consult the task execution process of a typical business process, to automatically analyze the business process tasks of various types of multidimensional spatial-temporal data processing interactions, to divide these business process tasks into simple subtasks, and to autonomously execute, transfer, and solve the subtasks; to convert various types of multidimensional spatiotemporal data into corresponding natural language descriptions based on the inference and understanding capabilities of the large-scale spatial-temporal vector model; to complete the interaction, analysis, and intergeneration of intelligent, body-controlled, multidimensional vector graphics and topic-specific graphic documents in the field of technology using simple inputs including speech and text; and to accurately capture the parameters in the multidimensional spatiotemporal data and automatically update the relevant multidimensional spatiotemporal data based on the inference capabilities of the large-scale spatiotemporal vector model and the processing capabilities of the intelligent body, thereby realizing the automatic generation and updating of the multidimensional vector graphics or topic-specific graphic documents; the process of constructing the large-scale spatial-temporal vector model includes: natural language description of the multidimensional spatial-temporal vector data, pre-training and fine-tuning of the large-scale spatial-temporal vector model, and understanding and analyzing the multidimensional spatial-temporal vector data; and wherein the multidimensional spatiotemporal vector large-scale model end uses the engineering domain data as a pre-training dataset to fine-tune the generic large-scale model, and wherein the generic large-scale model is pre-trained by extending the engineering domain lexicon using the pre-training dataset to enhance the ability of the generic large-scale model to understand and execute the engineering domain knowledge and data processing process; and wherein the processor is further configured to: describe the processing operations of the technical domain data, define the language of the processing operations of the technical domain data, use the defined processing operations of the technical domain data as a fine-tuning dataset to train the generic large-scale model, and enhance the ability of the generic large-scale model to understand and execute the processing operations of THE technical domain data; and wherein the natural language description of the multidimensional spatiotemporal vector data includes spatiotemporal vector data of various types of points, lines, surfaces and solids with continuous spatiotemporal (x, y, z, t) information and attribute information in various domains, including geographic information systems (GIS) and computer-aided mapping (CAD), and wherein, according to the requirements of training, understanding and processing the large-scale spatiotemporal vector model, the description transformation is performed in the form of natural language, and wherein the transformed natural language description contains all the information of the original spatiotemporal vector data, including the geometric type of the vector object, the geometric features, the display style, the coordinates of the reference point, and the relative geometric data based on the reference point.the temporal information and the attribute information; and wherein the pre-training and fine-tuning of the spatiotemporal vector large-scale model includes: creating a dataset that includes the natural language descriptions of the multidimensional spatiotemporal vector data and the vector object features, continuous pre-training and model fine-tuning on the generic large-scale model based on the pre-training dataset and the fine-tuning dataset, and forming a spatiotemporal vector large-scale model with the ability to understand the spatiotemporal vector data; and wherein understanding and analyzing the multidimensional spatiotemporal vector data includes: inputting spatiotemporal vector data described in natural language into the spatiotemporal vector macromodel, wherein the spatiotemporal vector macromodel outputs features of the understood vector object, and inputs some of the features of the vector object into the spatiotemporal vector macromodel, and wherein the spatiotemporal vector macromodel outputs complete spatiotemporal vector data that is understood or analyzed, and based on the output spatiotemporal features or data of the vector object, understanding and processing geometric features of the spatiotemporal vector data, understanding and processing attribute features,the differentiation and processing of spatial relationships and the differentiation and processing of spatiotemporal relationships are further complemented by the analytical capabilities of the spatiotemporal vector macromodel; and wherein understanding and processing geometric features of the spatiotemporal vector data includes: processing the geometric coordinates of the vector object itself; and wherein understanding and processing attribute features of the spatiotemporal vector data includes: processing the attribute information of the vector object itself; and wherein the distinguishing and processing of spatial relationships of the spatiotemporal vector data includes: distinguishing and processing the topological spatial relationship, the sequential spatial relationship and the metric spatial relationship between vector objects; and wherein the creation of a dataset comprising the natural language descriptions of the multidimensional spatiotemporal vector data and the vector object features, and continuous pretraining and model fine-tuning on the generic large-scale model based on the pretraining dataset and the fine-tuning dataset comprise the following: Step T1: Collecting a huge amount of spatiotemporal vector data, the corresponding natural language text descriptions of spatiotemporal vector data and the feature descriptions of the vector object, and forming a spatiotemporal vector training dataset, where the spatiotemporal vector data includes predefined geometric structure information of points, lines, surfaces and bodies, as well as attribute text information corresponding to the geometric bodies; Step T2: Perform pre-training with the generic large-scale model as a basis using the spatiotemporal vector training dataset, perform vector text sampling by random wandering as a training sample, input the sampled vector text as a training pattern into the generic large-scale model for fine-tuning the model; Step T3: Collect several types of question and answer pattern data, including vector data topological relationship, vector data attribute information, attribute description vector data, to form a spatiotemporal vector model fine-tuning dataset, and perform fine-tuning of the pre-trained generic large-scale model in step T2; and wherein the processor is further configured to: retrieve similar tasks from an engineering process library based on a task description, wherein the task scheduling mode includes scheduling without feedback and scheduling with feedback, and wherein the processor is further configured to report back the execution results from the environment, the user or a large-scale spatiotemporal vector model and to decompose the business process into a multitude of subtasks for execution; and wherein the processor is further configured such that the step of interacting, analyzing, and intergenerating the multidimensional vector graphics and the theme-specific graphic documents includes: processing and understanding multidimensional spatiotemporal data, generating a multidimensional spatiotemporal data processing intelligence body, and intergenerating multidimensional vector graphics and theme-specific graphic documents, which further includes: Step S1: Input of the multidimensional spatiotemporal data, conversion of the multidimensional spatiotemporal data into natural language descriptions that can be processed by the spatiotemporal vector macromodel, performance of the analysis and understanding of multidimensional spatiotemporal information vector data and temporal data in the multidimensional spatiotemporal data by artificial intelligence based on the reasoning and understanding capabilities of the spatiotemporal vector macromodel; Step S2: Input of user instructions through interactive means, including speech and text input; conversion of the user instructions described in natural language into formatted information system data based on the understanding, analysis, and processing capabilities of the spatiotemporal vector grand model; editing, querying, analyzing, and outputting the instructions to generate the multidimensional spatiotemporal data processing intelligence body; Step S3: based on the multidimensional spatiotemporal information processing intelligence body, using the fusion call and multiple execution rounds of the multidimensional spatiotemporal information processing intelligence body, realizing the automatic generation of automated and intelligent multidimensional vector graphics and theme-specific graphic documents and the bidirectional updating of intergenerational artificial intelligence applications, and realizing the intergeneration of multidimensional vector graphics and various types of multidimensional spatiotemporal data; and wherein the generation of the multidimensional vector graphics or theme-specific graphic documents includes: accurately capturing parameters in the multidimensional spatiotemporal vector data and subsequently merging these parameters with the template to generate the multidimensional vector graphics or theme-specific graphic documents. [2] Intergenerational artificial intelligence system according to claim 1, characterized by, that processing the geometric coordinates of the vector object itself includes modifying, editing, and manipulating shapes, sizes, and positions; wherein processing attribute information of the vector object itself includes modification, querying, analysis, and statistics; and wherein the topological spatial relationship refers to the association, proximity, inclusion, intersection, overlap, and separation relationship between the spatial objects; and wherein the sequential spatial relationship refers to the spatial order of the spatial objects or events in space, including front-back, left-right, top-bottom, and east-west-north-south orientation relationships; and wherein the metric spatial relationship refers to a distance or proximity relationship between spatial objects. [3] Intergenerational artificial intelligence system according to claim 1, characterized by, that the processor is further configured: based on the domain knowledge, business processes, and capabilities of the large-scale spatiotemporal vector model to understand the data and generate the text, to decompose and schedule the subtasks of interactive vector graphics business processes contained in various types of multidimensional spatiotemporal data, wherein the respective subtasks are the general logic processing or the spatial analysis operations of geographic information, executed independently with clear inputs and outputs; and wherein the respective subtasks are executed in an linked manner to complete a complex business process processing; when each task gathers information from the environment or receives information from memory, to provide the necessary data for each task and to store the process data during execution; wherein the multidimensional spatiotemporal information vector processing intelligence body gathers the data and experience and gradually completes self-evolution to provide support for the iterative capability of the spatiotemporal vector grand model; to execute each subtask planned by the spatiotemporal vector large-scale model into a specific result; wherein the task execution process relies on the inference capability of the spatiotemporal vector large-scale model, the spatial analysis and processing capability of the geographic information system and the processing capability of the computer-aided mapping system, and wherein the objects interacting in the execution process include sensors, model libraries, controllers and databases. [4] Intergenerational artificial intelligence system according to claim 1, characterized by that the several subtasks are linked together in a cascade or tree-like manner, with the subsequent subtasks being determined after completion of each subtask based on the results of the task execution. [5] Intergenerational artificial intelligence system according to claim 3, characterized by, that the processor is further configured to: store the information perceived by the environment, a record of task execution, the result of task execution, and to use the recorded memories to promote future actions, wherein the processor comprises stored data and operations; and wherein the stored data comprises a short-term memory of input information within a context window and a long-term memory of external vector memory, retrieved by a quick query. [6] Intergenerational artificial intelligence system according to claim 3, characterized by, that the processor remains configured to: use tools to complete the execution of individual tasks, wherein the tools include the large-scale spatiotemporal vector model itself and external tools, including algorithmic models, program compilations, databases, APIs, and wherein the processor continues to use spatial databases to perform some of the spatial analyses, uses models in the model library to complete specific tasks, and uses APIs to obtain real-time or historical data, and wherein the specific tasks are tasks corresponding to individual models in the model library. [7] Intergenerational artificial intelligence system according to claim 1, characterized by , that implementing the automatic generation of multidimensional vector graphics or theme-specific graphic documents includes the following: Artificial formulation of a template for the multidimensional vector graphics or the topic-specific graphic documents based on the content requirements of the multidimensional vector graphics and the topic-specific graphic documents, or automatic generation of a topic-specific graphic document template by the spatial-temporal vector macromodel based on the content requirements, wherein the graphics in the topic-specific graphic documents are in a multidimensional vector graphics format or in a raster format that has been converted from a multidimensional vector graphic; Extracting and generating various types of parameters in the template of multidimensional vector graphics or topic-specific graphic documents by querying and analyzing the information system or understanding and inferring the spatiotemporal vector macro-model, accurately capturing parameters in the multidimensional spatiotemporal vector data, merging the parameters with the template and generating the multidimensional vector graphics or topic-specific graphic documents. [8] Intergenerational artificial intelligence system according to claim 1, characterized by, that the intergeneration of the multidimensional vector graphics and the subject-specific graphic documents is based on the understanding and analysis capabilities of the spatiotemporal vector grand model, wherein, after the relevant changes to the spatiotemporal information caused by the technical domain data or the subject-specific graphic documents have been input, the spatiotemporal vector grand model generates the updated parameters, descriptions, or intelligence bodies of the multidimensional vector graphics in order to automatically update the data of the multidimensional vector graphics;and wherein, after the mapping or modification of the multidimensional vector graphics has resulted in a change to the multidimensional vector graphics, the modified multidimensional vector graphics are captured by the large-scale spatial-temporal vector model and the corresponding technical parameters or document descriptions are updated in order to update the relevant content of the topic-specific graphic documents.

Images