92results about How to "Realize automatic generation" patented technology

The invention discloses a method for dynamically constructing a UDF based on a large model. The method comprises the following steps: S1, constructing a multi-dimensional requirement cue word of a user-defined function according to characteristics and use requirements of a target platform; s2, the multi-dimensional requirement cue words are input into a pre-trained large language model for code generation, the large language model is subjected to multi-round iteration based on the multi-dimensional requirement cue words, and an optimal code sequence for achieving the user-defined function is output; s3, code style standardization is carried out on the optimal code sequence, and a compiler or an interpreter and a static analysis tool are respectively called to carry out grammar check and static analysis on the optimal code sequence, and constructing an optimization cue word according to an error log output by the compiler or the interpreter and a static analysis result output by the static analysis tool, and inputting the optimization cue word into the large language model to carry out iterative optimization on the optimal code sequence.

The present disclosure provides an intelligent scheduling method, a model training method and a device and system thereof. The method comprises: in the case that a target management node can utilize its own processing resources to schedule a kth rendering task, obtaining kth target state information corresponding to each slave node in N slave nodes contained in the next level of the target management node; the target management node is a jth management node in an ith level in the cluster; obtaining k-1th scheduling feedback information; inputting the kth target state information corresponding to each slave node contained in the next level of the target management node, the k-1th scheduling feedback information, and the task characteristics of the kth rendering task into a target neural network model to obtain a target scheduling strategy of the kth rendering task; and determining a target slave node from the N slave nodes according to the target scheduling strategy of the kth rendering task to allocate the kth rendering task to the target slave node.

The invention discloses a software test report generation method and device, a storage medium and computer equipment, and the method comprises the steps: responding to a test report generation instruction of target software for a target scene, and obtaining test index data of different data sources through a unified multi-data source adaptation layer interface, the different data sources comprise a Prometheus time sequence data source, a Grafana visual data source and a YAML test case data source; obtaining a test report generation template corresponding to the target scene, and determining an index association rule and a rendering type in the test report generation template; performing association analysis on the test index data of the different data sources on the basis of an index association rule, determining test conclusion data of the target software on the basis of an association analysis result, and rendering the test index data in the Grafana visual data source on the basis of a rendering type to obtain a rendered picture; and embedding the test index data, the test conclusion data and the rendering picture into a test report generation template to obtain a test report corresponding to the target software.

This application relates to the field of power distribution system technology, and in particular to a testing method and system for substation measurement and control devices based on neural networks. The method includes: constructing a full-link signal topology map of the substation; extracting the correlation features of each signal node from various types of configuration files; inputting the full-link signal topology map and correlation features into a pre-trained graph neural network, which matches signal node pairs to obtain matching point pairs, and generating a test sequence consisting of simulated excitation signals and expected response information based on the matching point pairs; triggering the corresponding simulated excitation signal according to the test sequence, acquiring feedback signals in real time, comparing the feedback signals with the expected response information, and determining the test results. This application effectively reduces the manual burden of testing substation measurement and control devices and improves test accuracy.

The application discloses a kind of based on visual perception's intelligent modeling generation system of moving head lamp, belong to stage lighting control technical field.The application includes: visual perception unit, obtains moving head light column three-dimensional information and establishes lamp space position database;Modeling case library is used to store the preset commonly used moving head light column modeling template;Space constraint analysis module is used to analyze the modeling space constraint that can be realized based on the actual space layout of on-site moving head lamp;AI modeling generation module is automatically generated based on lamp space distribution moving head light column basic modeling that adapts current lamp position layout;Modeling output interface is exported to console use.The application realizes the automatic generation of moving head lamp basic modeling based on on-site lamp position space perception for the first time, provides a series of position adaptation modeling scheme for lighting designer, and greatly improves programming efficiency.

The application relates to the technical field of digital human body models, in particular to a digital human body muscle model posture generation method and device. The method comprises the following steps: obtaining a tomographic image of an initial posture of a human body by using high-precision CT, and inversely modeling a geometric model of human body bones and muscles according to the tomographic image; calculating a first motion influence weight of each muscle node attached to a bone node according to each muscle node and the bone node; calculating a second motion influence weight of each muscle node according to the influence of adjacent muscle nodes on each muscle node; calculating a comprehensive weight of each muscle node according to the first motion influence weight and the second motion influence weight; and calculating a Laplace coordinate of a next moment of a current muscle node according to the comprehensive weight of the current muscle node, the position of the current muscle node and adjacent muscle nodes. The application generates a high-fidelity muscle geometric model by using a Laplace mesh deformation constraint technology.

The invention relates to the technical field of aircraft skin mirror image milling, in particular to a method for generating a mirror image milling track of a depression feature with a regular and protruding local contour of a skin. Selecting a closed programming contour boundary line, a skin outer surface and a cutter feeding control point, generating a cutter limit trajectory by the programming contour boundary line along the extension curved surface of the sinking bottom to the effective radius value of the cutter in the sinking area, performing finishing treatment on the limit trajectory, and generating a secondary limit trajectory along the extension curved surface of the sinking bottom to the distance value of the cutter in the sinking area; the secondary limit trajectory is ranked into an ordered line segment. And screening sub-limit trajectory segment boundaries under the condition that the sub-limit trajectory segment line segments are coplanar with the first line segment, identifying and classifying the sub-limit trajectory segment line segments into straight-edge end straight-edge protrusions, arc-edge end straight-edge protrusions, straight-edge end triangular protrusions or arc-edge end triangular protrusions, and correspondingly generating milling trajectory lines. And the milling trajectory of the remaining rectangular area is solved, the limit trajectory and the secondary limit trajectory are connected along the epitaxial curved surface, and finally a complete mirror milling trajectory is formed.

The application discloses a method, device and equipment for generating code and a storage medium, and belongs to the field of communication technology. The method comprises the following steps: reading message information corresponding to a plurality of messages in a DBC file, any message in the plurality of messages being used for transmitting at least one signal, and the message information of any message comprising a signal start bit and a signal bit width of the at least one signal; and acquiring a first code block of any message based on the signal start bit and the signal bit width of the at least one signal corresponding to the any message, the first code block comprising code lines corresponding to the at least one signal, and the code line of any signal being used for generating or identifying any signal in any message. The method automatically reads the DBC file to realize automatic generation of the code block, replaces manual operation of a user, improves the generation efficiency of the code block, and can avoid manual errors, thereby improving the accuracy of the generated code block.

The invention relates to the technical field of edge computing and automatic optimization algorithms, and discloses a multi-edge computing task unloading and scheduling hierarchical automatic optimization method based on large language model driving, which comprises the following steps: S1, constructing a computing system, acquiring task data and edge server data to form a data set, dividing the data set into a training set, an optimization set and a verification set; and S2, constructing a joint task unloading and scheduling optimization model taking the total completion time delay of the system as an optimization target, taking minimization of the total completion time delay of the system as an optimization target function, and taking the total completion time delay of the system obtained by the joint task unloading and scheduling optimization model as a fitness value. According to the method, efficient and automatic optimization of task unloading and scheduling problems in the multi-edge computing environment is achieved, the intelligent level and generalization ability of a computing system are improved through an algorithm evolution mechanism driven by a large language model, and the method has important theoretical significance and wide engineering application prospects.

The invention relates to a voice-based self-service wine selling device and method for self-service wine taking. The voice-based self-service wine selling device comprises a wine storage cavity, a data acquisition module, a wine amount calculation module, a two-dimensional code generation module, a locking module and a display module, a wine cover is arranged at the top of the wine storage cavity, a wine outlet module is arranged at the position, close to the bottom, of the wine storage cavity, a wine outlet is formed in the wine outlet module, a wine taking module is arranged below the wine outlet, the data acquisition module and the wine amount calculation module are arranged in the wine cover, and the data acquisition module is connected with the wine amount calculation module. The wine amount calculation module is connected with the two-dimensional code generation module, the two-dimensional code generation module is connected with the locking module, the locking module is connected with the wine taking module, and the two-dimensional code generation module is connected with the display module. By means of the device, automatic wine taking is achieved, cost calculation is automatically conducted, and a user conducts the whole process from wine taking to payment in a self-service mode.

The invention provides a multi-agent reward function automatic generation and optimization method and system, and relates to the technical field of artificial intelligence and reinforcement learning. According to the method, state description information of an environment is extracted through an environment context construction module, and then a plurality of reward function candidates are evaluated at the same time through a parallelized reward function evaluation module by utilizing a reward function code generated by a large language model; afterwards, statistical information in the training process is converted into structured natural language feedback through a reflection report generation module, so that the large language model can understand the training dynamics and improve the reward function in a targeted manner, and then context-based directional rewriting of the reward function is realized; and finally, through an iterative optimization process, generating a reward function which not only conforms to a task target but also has good training characteristics. According to the method, the code generation capability of the large language model is combined with the reinforcement learning training process, so that automatic generation and iterative optimization of the reward function are realized.

The application discloses a kind of meniscus data processing method and related equipment, method includes: based on sampling point set, by principal component analysis to obtain unit eigenvector and then construct to obtain meniscus main plane;According to basic data, by proportion demarcation identification to obtain meniscus front and back root;Sampling point set is projected to meniscus main plane, and then combined with meniscus front and back root to carry out polar angle partition, by polar radius weighted average screening representative point of each partition, construct to obtain continuous path curve;Meniscus front and back root are added to continuous path curve, obtain target path.The application constructs meniscus main plane by principal component analysis, and then combined with polar angle partition and polar radius weighted average, can realize the automatic generation of steel wire implantation path, and the path generated has good structural adaptability and repeatability, provides reliable support for the mechanical optimization and implantation safety of meniscus implant body, and can be widely applied in data processing technical field.

The invention discloses an airport pavement surface layer plate design method and system and a readable storage medium, and relates to the field of airport pavements, in the method, firstly, a block design parameter dictionary and design data are read to obtain a block design base, then point-line construction and seam line preprocessing are carried out by adopting a half-side data structure, and an airport pavement surface layer plate is obtained; and carrying out region segmentation and surface layer information assignment through an anticlockwise minimum region half grid method, then carrying out reinforcement design on underground pipelines, a filling and excavation boundary line passing region and a special-shaped region, and finally generating a three-dimensional model and assigning attribute information. According to the method, the associated configuration of the design parameters, the intelligent judgment of the reinforcement scheme and the seamless integration of the design data and the three-dimensional model are realized, and the design efficiency and accuracy are effectively improved.

This invention relates to the field of natural language processing and information extraction technology, and discloses a method for automatic generation and iterative optimization of annotation guidelines for event extraction. The method includes: generating initial annotation guidelines based on annotation samples using a large language model and compressing them into lightweight guidelines for predefined event types; employing a two-stage extraction strategy, first identifying event types using the lightweight guidelines, and then recalling complete guidelines based on the types for end-to-end extraction to obtain prediction results; identifying and clustering prediction errors based on a predefined set of event-layer and argument-layer error types to form error clusters; and generating update operations for the annotation guidelines using a large model for each error cluster, and completing iterative optimization after verification of effectiveness. This invention achieves automated generation and continuous evolution of annotation guidelines, and significantly improves the maintainability, generalization ability, and event extraction performance of the guidelines through structured representation and error clustering mechanisms.

Embodiments of the present application provide a kind of bus scheduling method, device, electronic equipment and computer readable storage medium, belong to intelligent transportation technical field.The method comprises: according to the first train chain set, static scheduling information and static basic information of bus operation, first driving plan information is calculated;According to the first driving plan information, execute driving plan;In the process of executing driving plan according to the first driving plan information, in the case of abnormal situation, obtain dynamic basic information of bus operation, historical passenger flow information and dynamic scheduling information;According to the second train chain set, dynamic scheduling information and historical passenger flow information, dynamic basic information of bus operation, second driving plan information is calculated according to static basic information of bus operation;And according to the second driving plan information, adjust and execute driving plan.The embodiments of the present application make the process of bus scheduling more intelligent, improve scheduling efficiency.

This invention discloses a method and system for generating multimedia children's picture books based on generative artificial intelligence. The method includes: acquiring a scanned image of a target physical picture book; identifying and extracting text information from the image; analyzing and extracting the subject, actions, and corresponding descriptive information; detecting the main object in the image and separating it from the background; generating a Stable-Diffusion model's introductory text based on the detected main object, the extracted subject, actions, and corresponding descriptive information, and combining it with a preset template; generating a dynamic image using a pre-trained Stable-Diffusion model; and completing the background image of the separated main object; inputting the dynamic image into a SwinBERT video summarization module to obtain a corresponding video description summary; and synthesizing corresponding narration, dialogue, or background sound based on the content of the dynamic image or pre-set content, thereby transforming a static picture book into a generative multimedia children's picture book. This method helps provide children with a personalized and interactive experience.

The application discloses a front-end page generation method and device, a program product and electronic equipment, and relates to the technical field of front ends, and the method comprises the following steps: analyzing annotation data of back-end code to obtain an analysis result, wherein the analysis result at least comprises a page identifier and component data of L page components, L is a positive integer, the page identifier is used for representing a function implemented by the back-end code, and the type of each page component is one of the following: a form type and a table type; determining a page template corresponding to the back-end code based on the page identifier, wherein the page template comprises at least one of the following: a form page sub-template and a table page sub-template; generating target front-end code based on the component data of the L page components and the page template; and generating a target front-end page corresponding to the back-end code based on the target front-end code. The application solves the technical problems of low generation efficiency and inconsistent page styles of front-end pages in the prior art.

The application discloses a network security threat assessment method, device and equipment and a readable storage medium, wherein the method comprises the following steps: determining a network security knowledge graph based on a network security knowledge system, acquiring network asset information corresponding to an existing network asset management system, determining a network asset topology graph based on attribute information of network assets in the network asset information, the first reachability relationship and the first connection relationship; generating a network security threat knowledge graph based on the alarm information, the network security knowledge graph and the network asset topology graph, and determining a threat assessment report corresponding to the network security threat knowledge graph. The application realizes automatic generation of the threat assessment report, greatly reduces the manual operation cost, improves the efficiency of outputting the threat assessment report, and improves the accuracy and comprehensiveness of the network security assessment through the network security knowledge graph and the network asset topology graph.

The invention belongs to the technical field of computers and artificial intelligence, particularly relates to an auxiliary teaching material generation method and system based on an artificial intelligence technology, and aims to solve the problems that existing teaching materials are low in production efficiency, insufficient in individuation degree and difficult to generate visual materials. The method comprises the steps of obtaining and analyzing a teaching material generation request to determine a teaching range requirement and a core knowledge point; according to the requirements and the core knowledge points, aggregating related contents from a plurality of resource libraries, and extracting and generating a hierarchical knowledge point structure by using an artificial intelligence large model; and generating at least one teaching material including an interactive animation, a knowledge card or a mind map based on the knowledge point structure. According to the invention, the teaching materials can be automatically, intelligently and individually generated, the lesson preparation efficiency is remarkably improved, the expression form of the teaching resources is enriched, and the teaching effect is improved.

The invention relates to a power report generation method and device, computer equipment and a readable storage medium, and relates to the technical field of power. The method comprises the following steps: extracting a report theme and data description information of a to-be-generated power report from a power report generation demand text; determining at least one initial analysis item matched with the report theme from a plurality of power report analysis items of a pre-constructed power report knowledge base, and obtaining and displaying an initial report outline; in response to an editing operation and a confirmation operation on the initial report outline, obtaining a target report outline; the target report outline comprises at least one target analysis item; obtaining target power data corresponding to each target analysis item from a power database based on the target report outline and the data description information; and according to the target report outline and the target power data corresponding to each target analysis item, generating and displaying a target power report. By adopting the method, the writing quality and writing efficiency of the power report can be improved.

The invention relates to a pipe fitting deployment method and device for a ventilation air conditioner air pipe system, and relates to the technical field of computers.The method comprises the steps that scene parameters of a target scene where the ventilation air conditioner air pipe system is to be deployed are obtained, according to the scene parameters, selecting a ventilation air-conditioning air duct required for deploying a ventilation air-conditioning air duct system for the target scene; geometric figure forming points of the ventilation air conditioner air pipe are determined, and an accessory placing coordinate system used for placing a three-way accessory model connected with the ventilation air conditioner air pipe is determined according to the geometric figure forming points; calling a three-way accessory model, and determining an accessory placement position of the three-way accessory model according to the accessory placement coordinate system and an original accessory coordinate system of the three-way accessory model; and the three-way accessory model is placed at the air pipe intersection of the ventilation air conditioner air pipe based on the accessory placement position to obtain a pipe fitting deployment result, and deployment of the ventilation air conditioner air pipe system is conducted on the target scene based on the pipe fitting deployment result. The deployment efficiency is improved.

This invention provides a method, apparatus, electronic device, and storage medium for generating unconditional CAD models based on face adjacency relationships. The method involves: obtaining the required number of CAD faces for a design; generating a topological structure based on a topological diffusion neural network model; inputting the topological structure into a geometric diffusion neural network to generate geometric hidden states; decoding the geometric hidden states to obtain explicit geometric point clouds; and performing CAD point cloud segmentation based on the explicit geometric point clouds and the topological structure to generate a CAD model. This invention achieves automatic generation of CAD models without conditional input by introducing a generation mechanism combining topological diffusion and geometric conditional diffusion, effectively reducing manual intervention and improving the consistency and precision of the generated results between the topological structure and geometric representation.

This application provides a model update method, apparatus, device, medium, and product. The method includes: acquiring target item features of a first item uploaded by a first user, the target item features including at least one of item usage status, item attribute parameters, and item identifier; performing a quality assessment on each target item feature using a quality assessment model to determine a first assessment result corresponding to each target item feature; acquiring target quality inspection information uploaded by a second user, the target quality inspection information including at least a second assessment result obtained by the second user on the quality assessment of the target item features in the second item; acquiring a first deviation between the first assessment result and the second assessment result, and updating the quality assessment model in reverse based on the first deviation. This method can improve the accuracy and timeliness of users determining the quality assessment results of target items online.

The invention provides a block chain-based data product registration management method, which comprises the following steps of: in response to manual auditing verification, marking a data product with metadata compliance as passing auditing; carrying out encryption processing on metadata of the data product; calling a block chain uploading interface, and packaging the encrypted data to a to-be-confirmed block; returning a unique evidence ID as an on-chain identifier; obtaining a complete hash value of the corresponding transaction based on the evidence ID; and storing the hash value and embedding the hash value in a certificate. According to the method, a transaction hash is embedded into a product registration certificate, a real-time verification mechanism on a chain is realized, and deep integration of a long safety chain and a certificate system is relied on: the transaction hash is generated by a long safety chain consensus mechanism and forms a binding relationship with metadata, and hash formats of other block chain systems cannot be matched with verification logic; certificate hash and on-chain metadata are quickly compared through a Changsafety link interface, and high-efficiency closed-loop verification is realized based on API design and consensus rules of the Changsafety link.

The application discloses a random test case generation method. The method comprises the following steps: generating at least one random factor, and acquiring a current test step in a plurality of type step libraries matched with a test object according to the value of each random factor; if the current test step satisfies a legality condition according to a generated test step set, the current test step is added to the generated test step set; returning to the operation of generating at least one random factor until the number of test steps in the generated test step set reaches a target number threshold; and constructing at least one random test case according to the generated test step set at the end of the cycle. Through the above technical scheme, the test case of the random scene can be automatically generated according to the test requirement, and the coverage of the test is effectively improved.

This application provides a script generation method and related apparatus. The method includes: receiving first user input from a terminal device; generating a story outline for a target script based on the first user input; generating a first plot node based on the story outline; determining the intentions of multiple characters; generating a second plot node for the target script based on the intentions of the first characters; and generating target text content associated with the intentions of the first characters. Using this embodiment, multiple plot nodes of a script can be generated based on user input, achieving automatic script generation, which is beneficial for improving script creation efficiency and saving script creation costs.

This application provides a method and apparatus for encrypted synchronization and trusted recovery of BIOS dual-image configuration, belonging to the field of server firmware configuration management. The method includes: in response to a change in configuration items of the BIOS primary image, the BMC obtains the changed configuration information, encrypts the configuration information, and constructs a structured backup package; stores the structured backup package in the BMC's local non-volatile storage area, and performs index management on the backup package stored in the local non-volatile storage area; in response to detecting a BIOS switch from the BIOS primary image to the BIOS backup image during boot, the BMC reads the backup package from the local non-volatile storage area and performs security verification on the backup package; after the security verification passes, the BMC loads the configuration information in the backup package into the BIOS backup image. The method and apparatus provided in this application can realize encrypted synchronization and trusted recovery of BIOS dual-image configuration.

This invention proposes a three-dimensional production line layout method, apparatus, electronic device, and readable storage medium. It configures corresponding workstation information for resource blocks in a two-dimensional production line layout diagram, wherein the workstation information consists of the workstation name and line name corresponding to the resource block in a production line resource library. It outputs a resource attribute table corresponding to the two-dimensional production line layout diagram, which includes the position coordinates of the resource blocks and the workstation information. Triggering a three-dimensional layout operation on the resource attribute table determines the resource model corresponding to the resource block in the production line resource library based on the workstation name and line name. The invention then generates a three-dimensional production line layout corresponding to the resource attribute table based on the resource model. During the conversion of the three-dimensional production line layout, the corresponding resource model is automatically retrieved from the production line resource library based on the workstation information, and then the three-dimensional production line layout is automatically generated based on the resource attribute table of the two-dimensional production line layout, achieving automated generation of the three-dimensional production line layout and improving its generation efficiency.

The application provides a data generation method, an electronic device, a medium and a product; wherein the method comprises: obtaining description information of a test requirement; obtaining a business parameter of the test requirement based on the description information; the business parameter comprises a target entity corresponding to the test requirement and a target entity relationship corresponding to the target entity; obtaining a target test case corresponding to the test requirement based on the target entity and the target entity relationship; and the test case can be automatically generated according to the test requirement.