26 results about "Processing" patented technology

The invention provides a malicious software API (Application Program Interface) call sequence detection method based on graph convolution. The method comprises the following steps: acquiring and recording API call sequence information of processes and sub-processes when a large number of software samples run; performing vectorization processing on the API calling sequence information; extracting aparameter relationship, a dependency relationship and a sequence relationship of the API function; establishing an API call graph; inputting the API call graph into a graph convolutional neural network for training to obtain a malicious software detection network model; collecting API calling sequence information of processes and sub-processes when the executable file to be detected runs; constructing an API call graph of the executable file to be detected, then inputting the API call graph of the executable file to be detected into the malicious software detection network model, If the output result of the malicious software detection network model is 1, indicating that the judgment result is malicious software; If the output result of the malicious software detection network model is 0,indicating that the judgment result is normal software.

The invention discloses a scene text recognition method based on man-machine cooperation, and the method comprises the following steps: S1, carrying out preliminary processing of an existing scene text data set, and selecting a pre-training data set, a training set and a test set from the scene text data set; S2, training an SEE network by using the pre-training data set to obtain a pre-training model; S3, predicting the unmarked training set by adopting a pre-training model, and dividing the unmarked training set into a Hard sample and an Easy sample according to the degree of confidence of aprediction label generated by the model for the unmarked training set; performing manual annotation on the Hard sample, performing pseudo annotation on the Easy sample by using a model, and then performing fine tuning on the scene text recognition model by using the annotated sample; and S4, repeating the step S3 until the performance of the model meets the expected requirements. According to themethod, the marking cost of the scene text data set can be reduced, and the character recognition model performance can be improved.

The invention discloses a software and hardware cooperative programmable forwarding system, method and device, and the method comprises the steps that a programmable forwarding configuration module inthe system schedules calculation resources based on a control and configuration instruction, and enables corresponding flow processing logic to be distributed to a software forwarding function moduleand a hardware forwarding function module; the hardware forwarding function module comprises a fixed function forwarding working mode and a software definition forwarding working mode, and is used for processing a fixed function data packet in the data stream and forwarding a processing result; and the software forwarding function module is composed of a software exchange process instance, processes a data packet based on scene change in the data stream based on the software exchange process instance, and feeds back a processing result to the hardware forwarding function module for forwarding. According to the system, the configuration of the hardware forwarding function module and the software forwarding function module can be realized based on the programmable forwarding configuration module, and the processing of a new network protocol can be realized based on the software forwarding function module, so that the forwarding efficiency is improved.

A method, system and apparatus for source code auto-suggestion, including receiving and processing source code from a source code repository, extracting one or more features from the source code received from the source code repository, extracting one or more features from the source code within a development environment, comparing the one or more features from the source code received from the source code repository with the one or more features extracted from the source code within the development environment, and providing a segment of source code from the received and processed source code repository according to a similarity to the source code within the development environment.

The invention provides a text classification data processing method and device, a storage medium and a program product, and relates to the field of data processing, in particular to the fields of natural language processing, big data, intelligent search, deep learning and the like. According to the specific implementation scheme, a plurality of initial label mining functions are constructed in advance, target parameters of the initial label mining functions are tested on a test set, and the label mining functions with the target parameters meeting training requirements of a text classification model are obtained and serve as final target mining functions; the target label mining function is used for generating the classification label of the initial text, a large amount of training data of the classification label with labels in a specific application scene can be obtained, a large amount of training data with accurate classification labels are provided for training of a text classification model, the cost of manual data labeling is greatly saved, and the efficiency of obtaining the training data is improved.

The invention provides a client log processing method, a client log processing device and computer equipment, which can determine an equipment identifier, a log reporting time period and a reported log type aiming at a target client according to application requirements in a log salvage scene, and send a log reporting instruction containing the information to target equipment with the equipment identifier, so that the target client of the target equipment responds to the log reporting instruction, screening a target log matched with the log reporting time period and the reported log type from logs stored in a database, and then performing segmentation processing on the target log according to a preset unit time length by utilizing the generation time of the target log, and after the to-be-reported log files are obtained, the to-be-reported log files are reported to the log server, and compared with a processing mode of directly reporting the screened target logs, the processing mode of fragmenting the target logs and then reporting the target logs improves the log reporting speed and success rate.

The invention relates to a product process modeling method for a smart factory. Archiving of a product production process flow is completed through a visual interface of a computer; a background control center obtains and analyzes a document; the production and the processing of products can be realized according to the production process flow by calling related equipment; in a production processflow archiving process, the setting of the production process flow of different products is realized by editing the process path, the processing program file and the production flow, the filing process of the production process flow can meet the design of the production flow of different products; in the modeling process of the production process, the dynamic configuration of a model is realized by dynamically modifying information such as a process path, a processing program file and a production flow; and for the same product, the production process can be edited and modified in the production process, so that the purpose of flexible production is achieved, and the purposes of quickly replacing a production line and switching the product process can also be achieved through dynamic configuration of the model.

The embodiment of the invention discloses an audio and video frame loss method and device, computer equipment and a storage medium. The method is applied to the technical field of video playing, and comprises the following steps: if a preset audio and video playing instruction is received, judging whether the current moment is within a preset sampling time period or not; if yes, calculating the processing frame rate of the sampling data; if not, calculating a current frame loss mark according to a preset frame loss mark and a preset frame loss interval; performing frame loss processing according to a current frame loss interval and a current frame loss mark determined by the processing frame rate and the preset playing frame rate; calculating a current processing frame rate; and taking the current processing frame rate and the current frame loss mark as a processing frame rate and a preset frame loss mark respectively, and returning to execute the step of calculating the current frame loss mark according to the preset frame loss mark and the preset frame loss interval until a preset ending instruction is received. According to the embodiment of the invention, smooth frame loss can be realized, and audio and video output synchronization is ensured.

A processing system and a method of operating a processing system are disclosed. The processing system increases a total memory space that is logically available for processors in a general purpose graphics processing unit (GPGPU) module by utilizing memory space in a corresponding external memory module, and by further utilizing a portion of memory space in a plurality of other external memory modules, to accommodate an amount of data of a terbyte size.

The invention provides a game model blending processing method and device and an electronic device. A graphical user interface of a game engine is displayed through the electronic device. The method comprises the steps that responding to a model obtaining operation for the graphical user interface, multiple to-be-blended game models are obtained; responding to a placement operation for the game models, model placement forms of the game models are determined; model resources corresponding to the game models are extracted through the game engine; and according to the model placement forms, the model resources corresponding to the game models are blended, and an overall game model corresponding to the model placement forms is obtained. Accordingly, automatic blending operation can be conducted on the game models in the game engine, and the blended overall model is obtained.

The invention provides a surface element file conversion method and device, computer equipment and a readable storage medium, and the surface element file conversion method comprises the steps: obtaining a first format surface element file, and extracting vector line data, annotation text data and road center line data from the first format surface element file; performing preset vector linearization topology processing on the vector line data to obtain processed vector line data; utilizing the processed vector line data to construct an initial surface element, and superposing the corresponding annotation text data and the road center line data to the initial surface element to obtain a new surface element; and defining and distinguishing the attribute of each new surface element according to the annotated text data and/or the road center line data, and generating a second format surface element file. According to the method, the conversion efficiency of the surface element file can be improved, and the attribute of the new surface element can be quickly defined and distinguished according to the extracted annotation text data and the road center line data, so that the accuracy of the converted surface element file is ensured.

The invention relates to the technical field of two-dimensional codes, and discloses a supervision evaluation method based on a two-dimensional code, which comprises a two-dimensional code supervision and evaluation method, and the two-dimensional code supervision and evaluation method comprises the following implementation steps: analyzing character codes to generate a two-dimensional code, scanning the two-dimensional code, identifying the character codes contained in the two-dimensional code, searching and analyzing whether the character codes exist in the database, if the character codes exist in the database, checking the two-dimensional code page, if the character codes do not exist in the database, adding the character codes contained in the two-dimensional code, and checking the two-dimensional code page if the character codes are successfully added, checking the two-dimensional code page and adding related evaluation. According to the supervision and evaluation method based on the two-dimensional code, the problems of high software development difficulty and poor feedback efficiency of an evaluation system due to the fact that a two-dimensional code use object is not clearly distinguished during background data processing in the prior art are solved.

The invention discloses a method for binding a software service to CPU cores. The method comprises the steps that each CPU core is bound with a preset number of threads; the utilization rate of each CPU core is counted; when a new task is generated, querying the utilization rate of each CPU core; and selecting the thread bound with the CPU core with the minimum utilization rate as a target thread, and adding the new task into a target task queue of the target thread. By applying the method, the service and the CPU core can be flexibly bound according to the utilization rate of the CPU core, the utilization rate of the CPU can be effectively improved, the overall processing efficiency of the IO service is improved, and storage read-write IO is optimized. The invention further discloses a device and equipment for binding the software service to the CPU core and a computer readable storage medium which all have the technical effects.

The invention discloses a data processing method. The method comprises the steps of obtaining multi-source data associated with a user; performing feature data extraction on the multi-source data to obtain feature data; the feature data is input to a pre-trained target prediction model, a target intention prediction result of the user is output, the target prediction model comprises a factorization machine module and an attention network module, the factorization machine module is used for carrying out low-order feature processing on the feature data, and the attention network module is used for carrying out attention network processing; the attention network module is used for performing high-order feature processing on the feature data; and performing intention processing operation according to the target intention prediction result. According to the method, the corresponding high-order features can be extracted from the multi-source data associated with the user, and intention prediction is performed in combination with the low-order features and the high-order features, so that the accuracy of user intention prediction of the target prediction model is improved.

The invention discloses an automatic design method for digital chip module link. The automatic design method adopts a module linking automatic design technology which comprises a module interface definition requirement, a module instantiation requirement, a file FileList file format requirement and the like; the invention has the characteristics of directness, flexibility, conciseness and expansibility, module link automation and batch script processing are achieved, module link automation is achieved, the integration efficiency is high, and the error probability is low; module interfaces are unified and standardized, modules can be integrated, and reusability is high; top-level designers and module designers can design in parallel, and the project research and development progress is accelerated.

The invention relates to a resource processing method and device, computer equipment and a storage medium. The method comprises the following steps: acquiring at least one resource structure; determining a binary-like tree of each resource structure; determining the number of first resources and second resources in each resource structure according to the binary-like tree; determining a target resource structure in the at least one resource structure by comparing the number of the first resources in each resource structure and comparing the number of the second resources in each resource structure; the number of the first resources and the number of the second resources in the target resource structure meet a preset condition. By adopting the method, the number of registers and adders in each DSP cascade structure can be counted based on the binary tree-like tree, the DSP cascade structure with less trigger and lookup table resource consumption is found, and the consumption of the trigger and lookup table resources is reduced.