28 results about "Process training" patented technology

The invention discloses a polyester knitted fabric surface antistatic treatment method and system, and relates to the technical field of polyester knitted fabrics, the method comprises the following steps: determining a treatment process scheme according to a target application scene, calling a sample process record, and constructing a standard process training set; constructing and training an antistatic performance prediction model based on the standard process training set in combination with a knowledge distillation mechanism; a cost evaluation factor is constructed, a process evaluation function is defined in combination with the antistatic performance prediction model and the cost evaluation factor, and the process evaluation function is associated with a penalty space; and by taking the process evaluation function as an optimization target, carrying out adaptive capacity expansion and shrinkage iterative optimization on the treatment process parameters, determining a target process parameter combination for the polyester knitted fabric according to an iterative optimization result, and carrying out surface antistatic treatment. Therefore, the technical effects of reducing the parameter adjustment cost, improving the adjustment efficiency and enhancing the anti-static treatment flexibility are achieved.

Certain aspects of the present disclosure provide techniques and apparatus for improved machine learning. In an example method, a first plurality of weights for a base model and a second plurality of weights for an adapter model associated with the base model are accessed. A quantized plurality of weights is generated based on the first plurality of weights, a first quantization scale for the first plurality of weights, and the second plurality of weights. A loss is generated based on processing training data using the quantized plurality of weights. An updated second plurality of weights is generated based on updating the second plurality of weights based on the loss. A machine learning model comprising quantized versions of the first plurality of weights and the updated second plurality of weights is deployed.

The method performed by an electronic device may include: outputting monitoring data of a target scenario among one or more predetermined scenarios related to simulation errors, in a production mode of a semiconductor process simulation based on test data input by a user; switching from the production mode of the semiconductor process simulation to a maintenance mode; and terminating the target scenario when a user input related to simulation maintenance satisfies a maintenance condition of the target scenario while in the maintenance mode.

The invention relates to the technical field of network security, and discloses a rebound shell detection response system and method, a medium and a product, and the method obtains tetrad information, a process parameter set and a process behavior chain in real time through Netlink, avoids missing detection caused by short process injection time, and reduces system resource occupation; a probability distribution table is constructed in the process training module to serve as a dynamic baseline, so that false alarms caused by baseline stiffness are reduced; in the engine analysis module, the characteristics and behaviors of the host layer process are analyzed in combination with a probability distribution table, and whether the host layer process is a rebound shell process or not is judged, so that the problems that traditional rule detection is easy to bypass and encrypted traffic cannot be identified are solved, and false alarms and missing alarms are reduced; the multi-dimensional depth defense principle is combined in the dynamic response module, the rebound shell process can be thoroughly eradicated, the problems that traditional detection response is delayed and blocking is not thorough are solved, and the real-time performance, the accuracy and the rapid blocking performance of rebound shell detection are improved.

This invention discloses a development and training method and system based on artificial intelligence in the logistics field, including the following steps: receiving input of logistics professional training requirements, preprocessing and standardizing the input content, completing the training environment configuration and resource initialization, and obtaining clear training objectives and resource configuration parameters; based on the determined training objectives, collecting relevant data of logistics scenarios. This invention deploys full-process logistics professional training by constructing an edge-cloud collaborative training architecture, breaking the limitations of traditional training equipment and scenarios, improving the flexibility and practicality of training, processing training data through a combination of automated annotation and manual verification to improve data annotation efficiency and accuracy, quickly constructing high-quality standardized logistics training datasets, and building AI models for logistics scenarios through hyperparameter optimization and incremental training mechanisms to improve model adaptability and recognition capabilities, and lower the model development threshold.

The invention provides a model training method and device, a task processing method and device, equipment, a medium and a product, and the model training method comprises the steps: processing training data according to a text generation model of a current training round to obtain a prediction text set and a sample label set corresponding to the training data, determining reward values corresponding to the prediction text set in a plurality of preset optimization dimensions; according to the reward value corresponding to each preset optimization dimension, determining an advantage value of each preset optimization dimension; according to the preset dimension weight of each preset optimization dimension and the advantage value of each preset optimization dimension, determining a dimension loss value of each preset optimization dimension; and training a text generation model according to the dimension loss values of the plurality of preset optimization dimensions to obtain a text generation model of the next training round. According to the embodiment of the invention, the training precision of the text generation model can be improved.

The invention discloses a self-adaptive promotion robot driving training method based on capability grading and task decomposition, and aims to overcome the defects that in the prior art, test taking is emphasized, the capability is light, feedback is extensive, and a path is rigid. According to the method, a pyramid type three-layer teaching framework is adopted, and basic vehicle control training, atomic operation fragment strengthening and comprehensive process training are sequentially carried out. The system collects operation data in real time, constructs a dynamic ability portrait based on indexes such as vehicle speed standard deviation, calculates a comprehensive score through a weighting formula, realizes self-adaptive promotion in combination with a threshold value, and supports personalized point location marking and flexible splitting training. According to the invention, transformation from test taking to ability is realized, training precision and efficiency are improved, individual demands of trainees are adapted, teaching resource configuration is optimized, and the method is suitable for intelligent driving training scenes.

The utility model provides an automatic rifle live ammunition simulation process training device, and relates to the field of simulated live ammunition training, four corners inside an installation assembly are respectively provided with an installation hole, the upper side and the lower side of the installation hole are communicated in two directions, and the rear side of a guide mechanism is fixedly connected with a shooting assembly. The problems that in the prior art, although some simulation training devices exist, the simulation training devices have obvious defects in the aspects of simulation of real spot shooting feeling, operation feedback, correction guidance and the like, and the high-precision and high-efficiency spot shooting training requirements cannot be met are solved. The rack assembly is arranged at the bottom end of the servo motor A and is in meshing transmission with an out-stroke gear arranged on an output shaft at the bottom end of the servo motor A, so that when the recoil force feedback training device is used, real recoil force feedback can be simulated through meshing transmission of the driving gear and the rack assembly to be used by trainees, and then the purpose of precise training is achieved.

This application discloses a model training method and a cloud management platform based on a cloud management platform, which can improve model training efficiency. The method includes: a tenant sending a model training request for a specific model to the cloud management platform, whereby the tenant sets the number of queries, keys, and values ​​for the model. Based on this request, the cloud management platform determines multiple computing nodes. If the number of computing nodes exceeds the number of keys, the cloud management platform decomposes the model into multiple sub-models and deploys them on these computing nodes. These computing nodes can use these sub-models to process training data, obtaining the queries, subkeys, and subvalues. Then, these computing nodes can fuse the subkeys and subvalues ​​to obtain the keys and values. Subsequently, these computing nodes can use the queries, keys, and values ​​to complete the subsequent training of the sub-models.

The invention discloses a heterogeneous edge calculation multi-teacher federated distillation method based on course learning, and aims to reduce the influence of statistical heterogeneity on model performance on the premise of protecting the heterogeneity of an edge device model. The method specifically comprises the following steps: S1, a client performs local training by using a private data set and global knowledge, and then transmits output obtained by reasoning on a public data set to a server; s2, the server evaluates the quality of the output uploaded by each client by taking a sample as a unit, and then performs weighted aggregation to generate comprehensive knowledge; and S3, the server follows the thought of course learning, dynamically trains a global model by using comprehensive knowledge, and then sends global knowledge output by the global model to the client. And repeating the above process, and training until the global model converges. Compared with a previous federated learning method, the method has the advantages that the performance on four data sets is obviously improved, and the method is generally suitable for various training tasks.

This invention discloses a modular VR / AR surgical teaching system and method based on a local large-scale model. The local large-scale model training module is deployed in a local computing environment, storing structured medical knowledge and dynamically generating teaching scripts and assessment question banks. A lightweight VR / AR hardware group collects multi-dimensional user operation data and provides multi-sensory feedback. A full-process training control module, with a built-in physical simulation engine and evaluation logic, constructs a virtual surgical scene, simulates tissue mechanical changes in real time, compares user operations with standard surgical procedure models and generates error correction guidance, and generates a quantitative ability assessment report based on the entire process data. A multi-scene data mapping interface adopts a modular architecture based on a standardized intermediate communication protocol, achieving low-coupling logical isolation and dynamic integration between functional modules through a unified data exchange format and abstract interface specifications. This application adopts an edge-side deployment strategy to protect patient privacy and ensure real-time force feedback and tissue deformation rendering in the VR / AR scene.

The present application relates to a computing system, a model training method and apparatus, and a product. The computing system relates to a computing unit, and the computing unit comprises: a main board, which is configured with a central processing unit (CPU); and a base board, which is connected to the main board by means of a first communication link, wherein the base board is configured with a plurality of accelerator cards, and the plurality of accelerator cards are connected to each other by means of a second communication link. The main board is used for splitting a training task of a target model into a plurality of concurrent model training tasks and releasing same to the plurality of accelerator cards, and processing training results of the plurality of accelerator cards, so as to obtain a trained target model. The plurality of accelerator cards are used for concurrently executing the respective model training tasks thereof, so as to obtain the training results. The computing system forms an elastically scalable computing system architecture by means of modular base-board design and interconnection, such that the computing power and bandwidth of the computing system can match model training tasks at different parameter scales.

An XR-based semiconductor manufacturing process training device according to an embodiment of the present invention is configured to determine set values for parameters based on user input data, identify the spec data to which the set values belong among at least one spec data corresponding to the parameters defined in each manufacturing process, assign weights corresponding to the spec data to the set values, and calculate and display final data based on the first motion information and the weighted set values.

Embodiments of the present application disclose a task processing method, device and equipment, a computer storage medium and a computer program product, which comprise: performing grouping processing on a to-be-processed training task for model training to obtain multiple groups of training subtasks, and assigning each group of training subtasks a corresponding service node; executing the multiple groups of training subtasks in a parallel processing manner through the multiple service nodes corresponding to each group of training subtasks; stopping execution of a target group of training subtasks in which a target service node is located when a fault occurs in the target service node during execution of the multiple groups of training subtasks; determining a backup service node, and executing unexecuted training subtasks in the target group of training subtasks through the backup service node and other service nodes; the other service nodes are service nodes corresponding to the target group of training subtasks except the target service node, and the problem of a large amount of waste of processor time is solved.

The invention provides a model training method of a text processing model and a text processing method.The model training method of the text processing model comprises the steps that a current training sample set is obtained, and the current training sample set comprises sample texts and text labels; inputting each sample text into a text processing model, obtaining a prediction result corresponding to each training sample output by the text processing model, and training the text processing model according to the text label and the prediction result; determining a training sample with a prediction result different from the text label as a to-be-processed training sample, and obtaining an error type and sample scene information corresponding to the to-be-processed training sample; generating an expanded training sample according to the training sample to be processed, the error type and the sample scene information, and adding the expanded training sample to a next training sample set; and continuing to train the text processing model based on the next training sample set. Through an iterative training mode, the model training efficiency and the model processing capability of the text processing model are improved.

The invention provides a source code-oriented fine-grained vulnerability detection method and system. The method comprises the following steps of: acquiring existing program source codes and vulnerability sensitive elements as training data; compiling a program source code in the training data into an intermediate representation, performing inter-process slicing on the code value flow graph by taking a central node containing vulnerability sensitive elements as a reference, and extracting information related to vulnerabilities in codes; performing vector representation on the generated slice sub-graph, and converting information of a graph structure into a format capable of being input into a vulnerability detection model; training a vulnerability detection model by using the processed training data; and extracting a slice sub-graph of the source code of the target program, generating node embedding, inputting the node embedding into the trained vulnerability detection model for detection, and predicting whether the slices of the target program contain vulnerabilities or not and the specific vulnerability positions. According to the method, high accuracy of vulnerability detection can be realized, vulnerabilities can be detected on fine grit, and vulnerability positions can be accurately positioned.

The invention provides a model generation method, a recommendation method and a related device. The method comprises the following steps: constructing a single upright post connecting operation process training data set; performing data treatment on the training data set in the operation process of the single upright post to obtain a treated data set; based on feature engineering, performing data feature extraction on the treated data set to obtain a single upright post connecting operation process target training data set, and determining relevance features of the extracted features; and based on the relevance features and the target training data set in the operation process of the single upright post, training to obtain an intelligent recommendation model of the full-time nodes of the operation process of the single upright post and the working parameters of the single upright post. The method is used for improving the accuracy and reliability of the intelligent recommendation model of the working procedure full-time nodes of the single connecting stand column and the working parameters of the single connecting stand column, so that the intelligent recommendation model can keep efficient and stable performance in different application scenes, and a solid foundation is provided for scientific and accurate drilling operation; and the drilling operation efficiency is further improved.

A computer-implemented integration management system for the operational onboarding of a personnel unit into an institutional IT environment, wherein the system includes the following: a storage unit that is operationally coupled to a processor unit and configured to store role definition records, operational evidence records, milestone evaluation records, and integration evaluation records; a display interface unit configured to display information on integration progress, milestone assessment results, and operational task plans on a display device; a network communication interface configured to exchange institutional system data with external computer systems, including human resources management systems, institutional platforms, and communication servers; a data interface unit for role configuration, configured to receive role definition records that include task specifications, regulatory compliance parameters, institutional resource identifiers, and operational milestone plans linked to the personnel unit; a processing unit executed by the processor unit for generating a role matrix, configured to generate a role requirement matrix record from the role definition records, wherein the role requirement matrix record includes task identifiers, responsible entity identifiers, deadline parameters, and verification check fields stored in memory; a processing unit for deriving deployment plans, configured to generate a multi-level operational deployment record from the role requirement matrix record, wherein the deployment record includes sequential operational phase records and corresponding task execution parameters; an identity provisioning control unit configured to generate digital identity records, including authentication data, institutional account identifiers, and platform access authorization records, which are linked to the personnel unit; a training data processing unit configured to store and process training module datasets that include training module identifiers, competency verification parameters, and records of training completion; a milestone evaluation unit configured to process operational evidence records, including task completion logs, institutional system usage logs, mentoring interaction logs, and training completion logs, to generate milestone evaluation scores; a processor for calculating the integration value, configured to calculate an integration value based on the milestone evaluation values, the training completion records, and the stored operational evidence records; and a decision control unit configured to compare the integration value with predefined threshold parameters and generate appropriate adjustment control signals for operational reinforcement measures.

The invention discloses a model updating method and a related product. The method comprises the steps of obtaining to-be-processed training data; the to-be-processed training data comprises unlabeled data and labeled data; performing clustering processing on unlabeled data in the to-be-processed training data to obtain multiple groups of unlabeled data clusters; generating a label for each group of label-free data clusters in the to-be-processed training data to obtain the training data after the first processing; carrying out expansion processing on data with labels in the training data processed for the first time to obtain the training data after expansion processing; and carrying out updating training on the model by utilizing the training data after expanding writing. According to the method, the dependence of traditional supervised learning on large-scale manual annotation is effectively relieved, and the annotation cost and the iteration period are remarkably reduced; meanwhile, by introducing a long-tail semantic mode in the label-free data, the generalization ability of the model for unseen expressions is enhanced, the recognition effect on low-frequency intentions is improved, the efficiency of generating new data in the model updating process is improved, and the cost is reduced.

Certain aspects of the present disclosure provide techniques and apparatus for improved machine learning. In an example method, a first plurality of weights for a base model and a second plurality of weights for an adapter model associated with the base model are accessed. A quantized plurality of weights is generated based on the first plurality of weights, a first quantization scale for the first plurality of weights, and the second plurality of weights. A loss is generated based on processing training data using the quantized plurality of weights. An updated second plurality of weights is generated based on updating the second plurality of weights based on the loss. A machine learning model comprising quantized versions of the first plurality of weights and the updated second plurality of weights is deployed.

The embodiment of the specification discloses a method for processing training data of a large model. The method comprises the following steps: obtaining an initial data set containing an inference process of the large model; determining, from the inference process, a condition required by the large model for inferring an output result in a model input of the initial data set; updating the determined condition in the model input, so that the model input no longer has the condition required by the large model for outputting the result, obtaining a training data set labeled as unable to output the result, and fine-tuning the large model to be adjusted based on the training data set. The method determines the necessary condition for the output result through the inference process of the large model, and then updates the corresponding condition in the model input based on the condition, so that the training sample set no longer has the condition for inferring the true output result. Thus, the large model can recognize the ability of the model input without the output result, and the "illusion" of the output result of the large model is avoided.

The invention relates to an evaluation method for aviation emergency simulation rescue comprehensive training, belongs to the technical field of emergency rescue task training evaluation, and solves the problem that in the prior art, aviation emergency rescue training lacks multi-role collaborative comprehensive evaluation. A task process training evaluation model is established and used for executing task completion condition evaluation, including individual completion condition evaluation and unit completion condition evaluation, and individual completion condition evaluation and unit completion condition evaluation are obtained; s2, establishing a task cooperative training evaluation model for obtaining a unit cooperative training evaluation value; and S3, obtaining task training parameters through VR training, providing the task training parameters to the established task process training evaluation model and the task cooperative training evaluation model, obtaining individual completion condition evaluation, unit completion condition evaluation and unit cooperative training evaluation values as evaluation results, and providing the evaluation results to an analysis process and a decision process.

The application relates to a multi-label learning method, device and equipment based on sample missing label enhancement, which comprises the following steps: acquiring a training data set of a missing label sample; pre-processing the training data set to obtain a processed training set with restored real labels; learning and aggregating the processed training set by using an algorithm adaptation strategy to obtain a multi-label learning classifier; taking the classifier as a label prediction model; and inputting a sample to be predicted into the label prediction model to obtain labels corresponding to the sample to be predicted. The method realizes label information enhancement by obtaining the processed training set with restored real labels; then the processed training set is induced by using the algorithm adaptation strategy to obtain a classifier considering the class imbalance problem in the processed training set; and the label prediction model is constructed based on the classifier to solve the multi-label class imbalance problem and improve the precision and accuracy of the predicted labels.

An example apparatus includes interface circuitry to obtain data; samples to train an AI-based model; machine readable instructions; and at least one programmable circuit to at least one of instantiate or execute the machine readable instructions to: transform the data samples into features; generate hash signatures for corresponding ones of the features; group the features into clusters based on the hash signatures; generate a filtered data set by filtering out features within a cluster of features having more than a threshold number of features; and train the AI-based model based on the filtered data set.

A system for training a segmentation model, comprising: an interface configured to allow a user to: upload and store training data in a storage device of a cloud-based network; provide access to the training data stored in the storage device; initiate a request to train a segmentation model; monitor training of the segmentation model; and download the trained segmentation model; and a computing infrastructure configured to: pre-process the training data using a first set of computing resources of the cloud-based network to obtain processed training data, and store the processed training data in the storage device; deploy a training application on a second set of computing resources of the cloud-based network to train the segmentation model based on the processed training data; provide access to monitor the training; and provide access to the trained segmentation model.

The invention provides a task scheduling optimization method and device, and relates to the technical field of data processing. A task scheduling optimization method comprises the following steps: S1, using a column generation algorithm to solve a process training reinforcement learning model of a sample task to obtain a column generation strategy model; s2, constructing a current target main problem and a current target sub-problem of a column generation algorithm according to the target task; s3, column generation iteration is carried out, and the relative decrease amplitude of the target function value and a newly generated column are obtained; s4, if a newly generated column with the check number smaller than 0 exists, inputting the relative decrease amplitude into the column generation strategy model to obtain a target column added in current iteration, updating the current target main problem and the current target sub-problem, and skipping to the step S3; and S5, if the newly generated column with the check number smaller than 0 does not exist, taking the solving result of the current target main problem as a scheduling optimization result. According to the method and the device, the decision variable optimization problem is efficiently solved.

The application relates to the technical field of computer vision, and discloses a tool detection method, a tool detection system and a storage medium. The method comprises the following steps: acquiring a data set; preprocessing the data set to obtain a training set and a verification set; performing data enhancement processing on the training set to obtain a processed training set; training an initial tool detection network model by using the processed training set to obtain a trained tool detection network model; optimizing parameters of the trained tool detection network model by using an improved QFocalLoss loss function to obtain an optimized tool detection model; performing data enhancement processing on the verification set to obtain a processed verification set; verifying the optimized tool detection model by using the processed verification set to obtain a detection result. The application can improve the robustness and accuracy of tool detection.