16 results about "Class Code" patented technology

The invention provides a camouflage object semantic segmentation method and device based on an adaptive candidate strategy, and a medium, and the technical scheme of the invention obtains the category code of a to-be-segmented image through a coarse feature extractor and a classifier, achieves the transmission of semantic category information, and enhances the semantic perception capability of a subsequent segmentation task. Initial prediction of a camouflage object is carried out through a target detector and an adaptive candidate strategy, an optimal target attention box is generated, and robust space guidance is provided for a core segmentation task; according to the method, category coding and a target attention box are combined, multi-source information is fused and reconstructed through a multi-guide feature fusion module and a multi-task perception decoder, and an accurate segmentation result is output.

This invention discloses an automated sports equipment management system and method based on machine learning, comprising the following steps: acquiring borrowing and returning trigger data, class code image acquisition, ball return area image acquisition, ultrasonic distance data, and equipment quantity data; inputting the class code image into an improved YOLOv2 ball borrowing and returning status recognition model deployed in a K210 vision module to obtain class recognition results; controlling the servo motor to release the ball based on the class recognition results and the number of remaining balls; opening the ball return cover and acquiring ball return detection data when returning the ball; generating a ball return confirmation result by combining ultrasonic distance changes and ball target recognition results; updating the number of borrowed balls in the class and the number of remaining balls in the equipment, and generating display, voice, and recording data. This invention achieves automated borrowing and returning management of ball sports equipment based on machine learning visual recognition and sensor linkage control technology.

Systems, methods, and other embodiments associated with a centralized product catalog with standardization from independent data sources are described. In one embodiment, a computing system is configured to access and download product information for a plurality of products that is maintained in a non-standard format. Metadata is identified and extracted. Feature vectors may be generated from the metadata and used to predict and assign class codes to products. The product attributes are extracted from each product, which include a non-standard product description. The non-standard product description is converted into a standardized text description. Standardized records are generated for each product record and a centralized database is generated with the standardized records.

The disclosure provides a camouflage object semantic segmentation method and device based on an adaptive candidate strategy, and a medium. The technical scheme of the disclosure obtains the class code of a to-be-segmented image through a coarse feature extractor and a classifier, realizes the transmission of semantic class information, and enhances the semantic perception ability of a subsequent segmentation task. An initial prediction of a camouflage object is performed through a target detector and an adaptive candidate strategy, and an optimal target attention box is generated to provide robust spatial guidance for a core segmentation task. The class code and the target attention box are combined, multi-source information is fused and reconstructed through a multi-guided feature fusion module and a multi-task perception decoder, and an accurate segmentation result is output.

The application discloses a software design and development method and a terminal, a detailed design document corresponding to an architecture design document is generated according to software requirements, a corresponding core class code is generated according to a core class diagram of the detailed design document, and an engineering code is obtained by directly developing based on the core class code by a developer; if the engineering code does not match the core class in the detailed design document, the core class diagram and the corresponding design document are re-generated, thereby greatly improving the consistency of software design and software development; if the engineering code matches the core class of the detailed design document, the engineering code is submitted to a code warehouse, the submitted document is ensured to be correct and reliable, and the design document is ensured to be well landed in the development stage.

The invention relates to a neural symbol multi-hop reasoning method and system for a domain knowledge graph. The method comprises the following steps: dividing a domain knowledge graph into multi-order sub-graphs, and performing symbol reasoning on each multi-order sub-graph according to a chained domain ontology to obtain an extended sub-graph; taking each multi-order sub-graph as a training sample, and coding instances in the multi-order sub-graphs to obtain class coding vectors and assertion coding vectors; constructing an assertion set reasoning network; taking assertions in the extended subgraph and the negative example set as a positive example and a negative example to construct a supervision label, and training an assertion set reasoning network by using a training sample and the supervision label; dividing a to-be-inferred domain knowledge graph into to-be-inferred multi-order sub-graphs, performing instance coding, inputting the to-be-inferred domain knowledge graph into the trained assertion set inference network, and obtaining corresponding extension sub-graphs according to the output extension type assertion and the established relation assertion; and combining the extended sub-graphs to output the extended domain knowledge graph. By adopting the method, a high-quality, strong-generalization and high-efficiency chain domain knowledge graph reasoning task can be realized.

The invention discloses a code abstract generation method fusing item-level code context, and belongs to the field of code abstract generation. Comprising the following steps: acquiring a code library to be subjected to abstract generation in a Java language, and acquiring a previous text, a next text and an abstract syntax tree language fragment of each segment of code in the code library; obtaining basic abstracts of various codes in the code library based on the trained model; obtaining context selection strategies of various codes based on the trained context selection strategy identification model; acquiring a context required by the class code from the context of the class code based on the context selection strategy of each segment of class code; and filling a preset prompt template with the basic abstract, the abstract syntax tree language segment, the required context and the metadata of each segment of class code, and inputting the prompt template filled with information into a large language model to obtain code abstracts of all class codes in the code library. According to the method, the code abstract with complete information and higher readability and accuracy can be generated.

The application relates to the field of data processing and communication technology, and discloses a standardized implementation method of a dynamic automatic translation device, which comprises the following steps: constructing and analyzing a standard configuration model, generating an attribute dependency matrix, performing loop detection and topological sorting, generating a linear ordered execution sequence, presetting a class code skeleton and a runtime context, extracting a logic code set according to the sequence, applying a code synthesis equation to generate a composite logic source code, calling an instant compiler and an isolated class loader to convert the source code into a dynamic translation instance and perform atomic reference replacement, and feeding device original data into the instance to make the device original data flow in a logical path in a one-way manner in the runtime context to complete translation. Through the conversion of the interpretive execution into the compiled sequential execution, the application eliminates runtime dependency lookup, cooperates with the isolated loading and atomic replacement mechanism, realizes the independent running and smooth hot updating of the device logic, and improves the efficiency and stability of the heterogeneous protocol adaptation.

The invention discloses a defect positioning method and system based on a large model standard operation process and a storage medium. The method comprises the following steps: step S1, defect understanding; performing byte code instrumentation on the failed test case, recording a method call chain and coverage class information, extracting a test tool class code and document metadata by adopting static analysis, enabling a large language model to integrate an error log, test output and an execution track for analysis, and generating a potential error cause hypothesis; s2, code library navigation; identifying all related methods from the whole code library by using a document enhanced hierarchical code search mechanism; and S3, defect verification. The positioning system and the storage medium are implemented based on the method. The method has the advantages of simple principle, higher defect positioning precision, higher efficiency and the like.

The invention provides a test case generation method and device, electronic equipment, a medium and a program product, relates to application of a large model in code testing, and can be applied to the technical field of artificial intelligence. The method comprises the steps of obtaining at least one target class code, and determining a target method of at least one to-be-generated unit test case from the target class code; for each target method, obtaining a corresponding target method code and historical test information, and constructing model input content; inputting the model input content into a large model to obtain a unit test case; and running the unit test case, obtaining test feedback of this round, and judging whether the unit test case meets a preset generation termination condition or not based on the test feedback of this round.

The invention discloses a zero-intrusion-oriented compiling period type conversion generation method. The method comprises the steps that a type conversion annotation in a source code is scanned and analyzed, a generation task is constructed for the type conversion annotation, and the generation task indicates annotated element information and a target type expression; for each generation task, respectively constructing a source type descriptor and a target type descriptor for the annotated element and the target type expression, the source type descriptor and the target type descriptor being based on the same data model; performing consistency verification on the source type descriptor and the target type descriptor in each generation task; and for the generation task which can be subjected to safety conversion according to the consistency verification judgment, a converter class code with a safe type is generated, safety enhancement logic based on a type conversion annotation is embedded in the converter class code, and safe type conversion is realized when the converter class code is called. According to the embodiment of the invention, on the premise of not changing service logic, the converter class code with the safe type can be automatically generated.

The application is suitable for the technical field of machine learning, and provides a code representation generation method for class code, which comprises the following steps: performing data reduction on a class-level source code file by adopting a class-level code data reduction strategy to obtain a first sequence and a second sequence after reduction; and performing representation fusion on a first vector representation sequence corresponding to the first sequence and a second vector representation sequence corresponding to the second sequence by adopting a representation fusion strategy based on an attention mechanism to obtain a code representation of the class-level source code file, so that the generation quality and effectiveness of the code representation of the class code are improved, and the model performance is improved.

The invention provides a database entity refreshing method and system and a storage medium. The method comprises the following steps: monitoring table structure change of a database to obtain structure metadata of each data table and a corresponding last change timestamp; each last change timestamp is compared with a corresponding refreshing timestamp of a local cache, and a target data table with a changed structure is judged based on comparison; triggering a dynamic code generation process for the target data table in response to the target data table; dynamically generating an entity class code and a data access layer component code corresponding to the target data table based on the target structure metadata of the target data table and a predefined code generation template; loading the entity class code and the data access layer component code into an application running environment through a dynamic class loading mechanism to replace or register the corresponding class definition; and updating the refresh timestamp information corresponding to the target data table in the local cache after dynamic class loading is completed. The database can be efficiently maintained at low cost.

The application discloses a kind of application reinforcement method and device, storage medium and electronic equipment. Among them, the method comprises: obtaining the code relationship diagram of target application program, wherein the code relationship diagram includes the reference relationship between each class in the bytecode assembly file corresponding to target application program;Based on the class code of each class, generate code retrieval strategy, wherein the code retrieval strategy includes a set of retrieval rules generated using the reference relationship between each class;Based on a set of retrieval rules, determine the code set to be processed from bytecode assembly file;The code in the code set to be processed is reinforced, and the reinforced target application program is obtained. The present application solves the technical problem of low efficiency in the application program reinforcement process in the related art.

Embodiments of the present application disclose an image processing method, device and equipment and medium, which are applied to the technical field of image processing. The method comprises: obtaining a target image containing a target object, an image feature vector and an object query vector; generating an object feature vector based on the object query vector and the image feature vector; obtaining a class code vector corresponding to each object class in an object class set; determining the similarity between the object feature vector and the class code vector corresponding to each object class, so as to determine the belonging object class of the target object from the object class set. The embodiments of the present application can improve the accuracy of object segmentation in an image.

The invention discloses a software cross-architecture migration workload evaluation method, which relates to the technical field of software, and is characterized by comprising the following steps of: matching an incompatible code segment to be transformed; dividing the incompatible to-be-transformed code segment into a quantum simple transformation type code and a quantum complex transformation type code according to the code segment modification quantum complexity, the associated file quantum entanglement depth, the modification scheme quantum superposition state probability and the quantum tunneling effect influence degree; the quantum basic workload of the class of quantum simplicity problems, the quantum learning efficiency factor of the quantum iteration period, the quantum learning curve curvature coefficient and the quantum iteration period number to which the class of quantum simplicity problems belong are calculated according to the quantum learning efficiency factor and the quantum learning curve curvature coefficient of the class of quantum simplicity problems; constructing a quantum ecological fragmentation coefficient and a quantum time expansion correction coefficient, and carrying out processing calculation to obtain the workload of a quantum simple transformation type code; the basic parameter set is set for the quantum complex transformation type codes, and the effect is that the workloads of different code transformation scenes can be accurately achieved.