7746 results about "Training methods" patented technology

The invention relates to the technical field of robot training. The robot self-adaptive training method based on reinforcement learning comprises the steps that task sub-target information is generated through a high-level strategy network, the task sub-target information is input into a low-level execution network, an action control instruction is generated according to the task sub-target information, interaction feedback information is collected in the execution process, and the action control instruction is sent to a robot through a robot. Calculating a reward value according to the interaction feedback information, carrying out association processing on the reward value and the scene complexity parameter, executing a dynamic reward shaping operation, generating an adjusted reward signal, generating a strategy model optimized by meta-learning based on the adjusted reward signal, loading the strategy model in a simulation environment, and carrying out dynamic reward shaping. A target strategy model optimized through simulation training is generated, the target strategy model is loaded to the robot, and the robot is controlled to execute task operation in the actual interaction scene. The method has the effect of realizing adaptive task learning of the robot in a multi-interaction scene.

The invention discloses a visual algorithm self-training method based on multi-agent collaborative optimization, and the method comprises the following steps: constructing a multi-agent system architecture which comprises a user interaction layer, an intelligent scheduling layer, an A2A protocol communication layer and a professional agent cluster layer; the user interaction layer analyzes a user task intention and generates an execution plan; the scheduling agent calls the professional agent to complete data processing, model construction, training, testing and deployment; a task process is coordinated through a standardized communication mechanism, and task execution is supported by combining an MCP tool set, a knowledge base module and a memory system; and when the task fails, automatically executing rescheduling operation, and finally outputting a self-training result. According to the method, the development efficiency, the self-adaptability and the intelligent level are remarkably improved, and the method is suitable for computer vision tasks such as industrial detection, intelligent security and protection and automatic driving.

The invention discloses a large model generation content traceability technology based on model copyright ID watermark embedding, which comprises the following steps that: firstly, a large model content generator compiles a dynamic watermark embedding process into an arithmetic circuit, and generates a watermark text and a proof by using a zero-knowledge proof algorithm; then the large model content generator publishes the text with the watermark and the proof together, and hides the copyright ID and the secret key; a large model content user verifies the proof and the watermarked text through a smart contract by adopting a zero-knowledge verification algorithm; after verification is passed, an adversarial sample corresponding to the text with the watermark is generated, an adversarial training method is used for optimizing the text with the watermark, and parameters of the Viterbi balance algorithm are dynamically adjusted and improved. According to the method, the concealment and the generation quality are balanced, high-capacity watermark embedding is realized, the tamper resistance and the accurate traceability of multi-model copyright disputes are improved, and the privacy protection intensity is improved.

The invention relates to a model training method and device. The method comprises the steps of obtaining a first training image set; performing first fine tuning training on the pre-trained artificial intelligence model by using the first training image set to obtain a preliminary optimization model; acquiring a second training image set; constructing a composite reward function based on the second training image set; wherein the composite reward function is used for performing multi-dimensional quantitative evaluation on the quality output by the model; and performing second fine tuning training on the preliminary optimization model based on the second training image set and the composite reward function to obtain a final optimization model. Therefore, a two-stage differential fine-tuning model strategy is realized, so that the output result of the final optimization model is highly matched with the expectation of a real scene, the problem of insufficient model practicability is fundamentally solved, and the reliability and value of the model after deployment are greatly improved.

The invention discloses a brain tumor image diagnosis large model pre-training method, diagnosis method and system, and the method comprises the steps: obtaining the image data of a brain tumor patient and a corresponding diagnosis text, and the image data comprises a plurality of sequences; constructing a visual unified model, carrying out complete sequence standard training and missing sequence distillation training on the visual unified model by utilizing the image data, learning unified visual representation of the image data of any sequence combination, and taking characteristics of the complete sequence image data as teacher characteristics in the missing sequence distillation training; the features of the missing sequence image data serve as student features, and distribution alignment of the student features and the teacher features in the feature space is restrained; constructing a visual language model, taking unified visual representation output by the trained visual unified model as input based on a multi-task target, and training the visual language model in combination with the diagnosis text; the trained model can effectively process the sequence missing condition, and the accuracy and robustness of diagnosis are improved.

The invention discloses an intelligent digital human training method and system based on multi-modal interaction, and belongs to the technical field of semantic indexing.The method specifically comprises the steps that voice, vision and text data are analyzed and converted into high-dimensional feature vectors through a modal exclusive encoder, the high-dimensional feature vectors are projected to a unified semantic space through a cross-modal semantic mapping model, and the high-dimensional feature vectors are obtained; generating a semantic primitive containing a modal identifier, a core semantic tag and a feature weight; semantic primitives are used as nodes, directed edges and edge weight table association strength are established based on semantic similarity, typical scene node connection weights are strengthened, and a mesh map containing intra-modal hierarchy and inter-modal cross association is formed; constructing a double-layer index on the basis of the mesh map; semantic primitives are extracted from newly added data, the position of a new node in an association graph is determined through a graph matching algorithm, an association edge with an existing node is automatically established, and a lower-layer modal exclusive index is synchronously updated.

The invention discloses a medical report generation method, a model training method, equipment and a medium, and the model training method comprises the steps: constructing a medical report generation model framework which comprises a global semantic collaborative multi-modal enhancement module, a visual encoder, a text encoder, a medical insight analyzer and an LLM decoder; wherein the global semantic collaborative multi-modal enhancement module respectively enhances a medical image and a medical report by utilizing a selected image enhancement strategy and a text enhancement strategy, and the medical insight analyzer comprises a fine-grained structure learning device and a global context guide learning device which are connected in sequence so as to enhance the cross-modal alignment capability; and performing intelligent collaborative optimization by taking a strategy set formed by an image enhancement strategy and a text enhancement strategy and architecture configuration parameters of the medical insight analyzer as optimization targets to obtain an optimal medical report generation model. The medical report generation performance can be effectively improved.

The invention provides a robot motion control model training method, device and equipment based on deep reinforcement learning, a robot and a medium, and relates to the technical field of robots. The method comprises the following steps: acquiring a first potential vector obtained after a student encoder encodes robot body observation data, and a second potential vector obtained after a teacher encoder encodes privilege observation data; based on the current training step number and a preset probability function, calculating a sampling probability for controlling a fusion proportion of the first potential vector and the second potential vector; fusing the first potential vector and the second potential vector based on the sampling probability to generate a third potential vector, and inputting the third potential vector into a strategy network; and updating the parameters of the policy network based on the value estimation of the current state output by the value network and the action policy output by the policy network. According to the method, updating oscillation caused by sudden change of input distribution in the training process of the strategy network can be avoided, the training efficiency is improved, and the training cost is reduced.

The invention discloses a federal large model knowledge collaborative training method supporting multi-modal heterogeneous clients, which comprises the following steps: each client receives a model initialization parameter issued by a central server, and applies adaptive differential privacy noise to independently train a heterogeneous lightweight model based on local private data; updating the model to which the noise is applied and uploading a modal identifier of the model to a central server side; after model updating and modal identification of each client are received, based on a modal perception weighted consensus fusion mechanism, knowledge of each client is fused to update a global large model; and the central server side issues the updated presentation layer parameters of the global large model to the client side for initialization of the next round of local training. According to the method, on the premise that a public data set or specific task setting is not needed, comprehensive compatibility of data isomerism, client dynamic participation, model diversity and privacy protection requirements is achieved, and the adaptability, stability and knowledge utilization efficiency of large model federation training are remarkably improved.

The invention discloses a large language model enhanced reinforcement learning training method for key automatic driving scene stable decision, and the method comprises the steps: constructing an LLM-based vehicle decision intelligent agent, and providing a high-level guidance strategy through a chain thinking technology and a knowledge experience pool; then, jointly constructing a dynamic intervention mechanism based on the environmental risk sign and the decision boundary parameter, thereby determining an LLM guidance opportunity; finally, through expert guidance algorithm design, the prior knowledge of the LLM is effectively integrated into the deep reinforcement learning strategy network by adopting an adaptive exploration-imitation fusion loss function. The DRL model is better in performance in a vehicle driving decision-making task and shows higher generalization ability in various scenes, and the robustness, the sample efficiency and the overall generalization level of the DRL model can be remarkably improved.

The invention discloses a reinforcement learning training method and system for relieving illusion of a multi-modal large model, and belongs to the field of reinforcement learning training of a multi-modal large language model. Firstly, a planning and visual description generation step is introduced in an early stage to guide a model to perform structured reasoning, then a grouping relative strategy optimization algorithm is used, reward values are calculated for multiple candidate responses generated by the model after cold start, and particularly, a visual perception reward mechanism is set. The reward mechanism evaluates the consistency of the generated text description and the visual information by using an external large language model. Then, based on a vision description attention score advantage distribution method, learning of the model on key vision signals is dynamically enhanced, and the perception ability of the model on the vision signals is improved; and finally, the perception and reasoning performance of the model is further improved by adopting multiple rounds of rejection sampling and supervised fine tuning. The scheme does not depend on a model architecture, the extra overhead is small, the illusion problem caused by early image-text inconsistency is effectively solved, and the accuracy and the reliability are improved.

The invention discloses a machine learning assisted polyethylene reaction performance prediction model training method, prediction method and device. The method comprises the following steps: acquiring a training set; screening feature items used for model training; obtaining a gradient boosting regression model for catalytic activity, a gradient boosting regression model for molecular weight and a gradient boosting regression model for molecular weight distribution; extracting feature items for model training from the data of the training set so as to obtain feature vectors; and respectively inputting the feature vectors into each model so as to train each model, thereby respectively obtaining hyper-parameters of the trained gradient-boosted regression model for catalytic activity, hyper-parameters of the trained gradient-boosted regression model for molecular weight and hyper-parameters of the trained gradient-boosted regression model for molecular weight distribution. According to the method, a model relationship between input characteristics and polymerization results (including catalytic activity, molecular weight, molecular weight distribution and the like) is established through training set learning.

The invention provides a reward model training method and device, a strategy model training method and device, electronic equipment and a storage medium, and belongs to the technical field of artificial intelligence, and the method comprises the steps: obtaining a preference data pair which comprises a preferred response and a non-preferred response generated for the same prompt word, and each of the preferred response and the non-preferred response is composed of a plurality of text unit sequences; inputting each text unit sequence into a to-be-trained reward model to obtain a predicted reward value; and calculating the total training loss according to the predicted reward value, and updating the model parameters of the to-be-trained reward model. According to the method, the response text is subjected to serialized splitting, and the preference data composed of the preferred response and the non-preferred response is introduced for comparative learning, so that the target of model training is no longer to evaluate the absolute quality of a single response, but to identify a key text unit which causes one response to be superior to the other response; the fine-grained evaluation of the response content is realized, and the evaluation accuracy of the reward model and the identification capability of complex user preferences are effectively improved.

The invention discloses a steel coil end face defect detection method and system, a training method and electronic equipment, and the steel coil end face defect detection method comprises the steps: driving a two-dimensional image and three-dimensional point cloud collection equipment to synchronously scan the end face of a steel coil through a movement mechanism, and obtaining registered RGB-D multi-modal data; inputting the data into a special detection model, extracting surface texture and geometric structure features, performing dynamic weighted fusion through channel splicing and a cross-modal attention mechanism, and outputting a suspected defect result containing defect types, positions, three-dimensional depth information and confidence; and carrying out geometric feature consistency verification on the suspected defect by combining the original point cloud, and finally determining a real defect. According to the model, deep synergy of texture and geometric features is innovatively realized, the recognition accuracy and robustness of complex defects such as micro cracks and recesses under reflection interference are remarkably improved, and meanwhile, the detection efficiency is guaranteed.

The invention provides an intelligent teaching-assistant system base LLM training method for a drilling simulator, and belongs to the technical field of petroleum drilling large language model training, and the method comprises the steps: obtaining the simulation data of the drilling simulator based on an intelligent teaching-assistant system base large language model, and carrying out the field self-adaptive pre-training; performing fine adjustment on the original weight by utilizing low-rank self-adaption, and performing operation steps and fault diagnosis generation; the method comprises the following steps: constructing an enterprise private knowledge base, training a large language model by utilizing retrieval enhancement generation, and obtaining a trained verification header through explicit supervision; a direct preference optimization loss function is defined, a direct preference optimization loss function of retrieval perception is obtained in combination with the verification head, the large language model is optimized, and large language model training is completed; according to the invention, an external professional knowledge base can be fused, an equipment operation mechanism can be understood, an intelligent assistant architecture with teaching guidance and error diagnosis capabilities is provided, and the intelligent level of the drilling simulator is improved.

The invention provides a handwritten mathematical expression recognition model training method, a handwritten mathematical expression recognition method and a handwritten mathematical expression recognition device, a LaTeX sequence is automatically extracted and enhanced from an academic PDF file, a large-scale and high-reality handwritten formula image data set is generated through rendering and material conversion, and then an end-to-end recognition model is trained. According to the scheme, the core bottleneck of training data scarcity and insufficient diversity caused by high manual labeling cost is fundamentally solved; according to the method, the generalization ability and robustness of the model in a real scene are effectively improved by synthesizing a vivid handwritten image; and finally, the identification model realizes a remarkable breakthrough in performance when processing a complex and long sequence formula, and a set of complete and extensible engineering solution is provided for large-scale application of the technology.

The invention discloses a whole-process vascular interventional operation training method and system based on a virtual reality technology, and the method comprises the steps: firstly, developing a vascular interventional mechanics and collision simulation module, and achieving the deformation simulation of a blood vessel and a guide wire through the combination of position dynamics, a region shape matching technology and a Cosseerat elastic rod model, the collision between the blood vessel and the guide wire is treated by adopting a wide-to-narrow three-layer progressive collision detection method; secondly, constructing an immersive virtual operation scene through model construction and rendering optimization; visual angle roaming, instrument operation and interface interaction are fused, and multi-mode interaction control driving intervention training is achieved; finally, based on a'module-event 'dual-drive framework, operation tasks such as anesthesia, puncture, radiography, intervention, stent release and hemostasis are sequentially connected in series through a unified state machine, and whole-process training is achieved. The accuracy and stability of blood vessel and guide wire simulation are improved, the training immersion sense is enhanced, interventional operation whole-process training is achieved, and high teaching and practical value is achieved.

A semantic segmentation model training method and apparatus, an electronic device and a storage medium are provided. The semantic segmentation model training method includes: acquiring a sample image, and extracting visual image features corresponding to the sample image by a semantic segmentation model to be trained; processing the sample image to obtain a text image feature corresponding to the sample image, the text image feature being an image feature generated from language description text for the sample image; fusing the visual image features with the text image feature to obtain multimodal features, and performing image segmentation prediction based on the multimodal features to obtain a target loss; and training the semantic segmentation model to be trained based on the target loss to obtain a target semantic segmentation model.

The invention relates to the technical field of medical simulation training, in particular to an ultrasonic operation intelligent training method and device based on multiple modalities, and the method comprises the steps: obtaining real-time six-degree-of-freedom pose data of an ultrasonic probe held by an operator, and generating an ultrasonic image in real time through a first deep learning model in cooperation with scene parameters; wherein the first deep learning model is trained to learn and establish a continuous mapping relation from an ultrasonic probe pose space to an ultrasonic image space; performing section classification on the ultrasound image by using a second deep learning model, and determining deviation information for a non-standard section; and based on the real-time six-degree-of-freedom pose data and a division result of a preset standard section and an ultrasonic probe pose, generating visual guide information for guiding an operator to adjust the probe pose, and displaying the visual guide information. In this way, the problems that an existing virtual training image is discontinuous and guiding is inaccurate are solved, and meanwhile the training efficiency and the reality sense of ultrasonic operation are remarkably improved.

This application discloses a model training method, a power prediction method, and a device. A model includes at least a first model and a second model. The method includes: obtaining a dataset including historical power data and historical meteorological data in preset duration; determining a weight of a loss function of each model based on the dataset; constructing a target loss function, where the first model and the second model may be used to represent input data, and data represented by the first model is not completely the same as data represented by the second model; and finally training the models based on the target loss function. More complete information is obtained from different modal data and different time scales based on a currently available data source, to optimize the learning model and achieve higher prediction precision.

A semantic segmentation model training method and apparatus, an electronic device and a storage medium are provided. The semantic segmentation model training method includes: acquiring a teacher semantic segmentation model that is pre-trained, the teacher semantic segmentation model including a first teacher network and a second teacher network, the first teacher network having structural characteristics of low depth and high width, and the second teacher network having structural characteristics of high depth and low width; processing a sample image based on the teacher semantic segmentation model to obtain a first segmentation map and a second segmentation map; and training a student semantic segmentation model that is lightweight according to the sample image, the first segmentation map and the second segmentation map, so as to obtain a target semantic segmentation model.

The invention relates to the technical field of power dispatching, and discloses a knowledge distillation-based low-resource power big language big model training method, system and device and a medium, and the method comprises the steps: obtaining accident case data of the power industry, carrying out the problem construction and task setting, introducing a quality evaluation mechanism, carrying out the refusal sampling through a language model, and carrying out the training of a low-resource power big language big model. Generating a distillation data set for model distillation; introducing a LoRA module into the student model for fine tuning, constructing multi-source heterogeneous fine tuning data, and setting a training strategy to optimize the performance of the model; and performing training by adopting reinforcement learning, introducing language consistency rewards until the reinforcement learning achieves convergence on the reasoning task, and generating a final language large model. According to the method, through knowledge distillation and reinforcement learning, the deep knowledge and the reasoning ability of the super-large model are successfully migrated to the small model, so that the parameter quantity of the finally deployed model is greatly reduced, and the computing power resource required by reasoning is sharply reduced.

The invention discloses a power industry large model continuous pre-training method and system based on dynamic self-constraint, and belongs to the technical field of artificial intelligence, and the method comprises the steps: obtaining industry pre-training corpora and instruction training corpora, dynamically adjusting the mixing ratio of the industry pre-training corpora and the instruction training corpora through a curriculum-type strategy, and obtaining an industry pre-training corpora and an instruction training corpora; obtaining a dynamic mixed data set; configuring a reference model based on the dynamic mixed data set, and training a target power industry large model by adopting a differential loss function and the reference model for different types of data in the dynamic mixed data set; according to the differential loss function, self-adaptive KL divergence is calculated according to inter-partition optimization logic, and the self-adaptive KL divergence is adopted to construct a loss function; and obtaining the probability of the reference model through an online reasoning framework, and enabling the training to be continuously carried out based on the probability of the reference model. According to the method, the knowledge conflict problem in professional field training is effectively solved, and the generality of the model is kept while the professional property of the power field is improved.

The invention provides a difficult case question and answer sample acquisition method, a model training method and related equipment. The difficult case question and answer sample obtaining method comprises the steps that a to-be-corrected question and answer sample is obtained, the to-be-corrected question and answer sample comprises a preset question, an annotated answer and a first reasoning link comprising a reasoning answer, and the reasoning answer included in the first reasoning link does not conform to the annotated answer; the to-be-corrected question and answer sample is input into a second language model, so that the second language model outputs first reflection content, and the first reflection content comprises an error point in the first reasoning link and a correction thought for the error point; inputting the preset question, the first reasoning link and the first reflection content into a second language model, so that the second language model outputs a second reasoning link including the reasoning answer; and if the inference answer included in the second inference link is consistent with the labeled answer, generating a difficult case question and answer sample based on the preset question, the labeled answer, the first inference link, the first reflection content and the second inference link.

The invention belongs to the technical field of customer service, and particularly relates to an intelligent customer service practical training system and method, and the system comprises a business scene digital reconstruction layer which is used for constructing a business training scene to form a training scene library, managing the business training scene library, and reconstructing a training scene according to scene configuration data; the AI simulation interaction core layer is used for receiving and recognizing voice and text input of students and generating behavior strategies of AI partner training according to input content on the basis of a current training scene; the training execution layer is used for generating scene configuration data according to a student selection instruction and sending the scene configuration data to the business scene digital reconstruction layer, and the scene configuration data content comprises a training scene, training difficulty and role setting; the training difficulty is dynamically adjusted according to the real-time performance of the trainee; and the quantitative evaluation layer is used for performing multi-dimensional analysis on the training data and generating a quantitative evaluation result and an improvement suggestion. The problems that existing customer service training is low in efficiency and high in cost are solved.

The invention discloses a multi-dimensional training method and device for a support vector machine, electronic equipment and a computer readable storage medium, and the method comprises the steps: carrying out the discretization of a training sample data set, and obtaining a discretized data set, the discretized data set comprises a plurality of different attributes, and each attribute corresponds to a plurality of feature vectors; calculating a classification contribution parameter of each attribute feature vector to obtain a plurality of classification contribution parameters; performing data mapping on the plurality of classification contribution degrees by using a kernel function to obtain a target function; and optimizing and training the objective function by using a gradient descent algorithm to obtain a support vector machine model. According to the method, different dimension data are mapped through the kernel function, the data gain weight can be determined, the linear separable effect of the mapped dimension data can be improved, and then the classification and analysis capability of the SVM can be improved.

The embodiment of the invention provides a model training method, a facial expression recognition method and device and electronic equipment, and relates to the technical field of video processing. The model training method comprises the following steps: acquiring a sample video and a first sample label; extracting a time feature and a space feature of the sample video by using a time-space feature extraction network in the facial expression recognition model of the initial structure; calculating an attention weight representing the correlation between the spatial feature and the time feature by using a mapping network; performing weighted aggregation on the time features by using the attention weight to obtain fused spatio-temporal features; inputting the fused spatial-temporal features into a classification network to obtain a first recognition result; and performing model training based on the difference between the first recognition result and the first sample label to obtain a trained facial expression recognition model with higher accuracy.

The embodiment of the invention provides a distributed model training method and device, a server and a storage medium, and relates to the technical field of distributed model training. Determining a target sub-model according to the training dependency relationship among the sub-models, and determining a target computing node according to a resource configuration parameter and a model training parameter of the target sub-model and the resource state of each computing node; creating a plurality of training tasks corresponding to the target sub-model according to the resource configuration parameters and the model training parameters of the target sub-model, allocating each training task to each target computing node, and controlling each target computing node to perform model training; and under the condition that the target sub-model is successfully trained, determining the next to-be-trained sub-model as a new target sub-model, and carrying out model training on the new target sub-model until all the sub-models are successfully trained, so that the flexibility and adaptability of resource allocation in the training process can be improved, and the training efficiency is improved. The problems of long resource occupation period, insufficient utilization and scheduling stiffness are solved.

The invention discloses a large model training method based on a cultural tourism scene. According to the specific implementation scheme, vertical corpora are extracted from a database, an initial data set is constructed, entities and relationships are extracted from the initial data set, and a text tourism knowledge graph is constructed; generating a question and answer pair based on manual annotation and the travel knowledge graph, inputting the question and answer pair into the reward model RM to output a quality score, and realizing initialization training of the reward model RM; improving interaction of the question-answer pairs by utilizing a knowledge verification large model to form new question-answer pairs; scoring the new question and answer pair by using the reward model RM which completes initialization training, filtering low-score samples, and taking high-score samples obtained after filtering as training samples of reinforcement learning; and using the training sample pair of reinforcement learning to finely adjust the reward model RM, and updating the parameters of the reward model RM based on the KL divergence. According to the invention, the professional degree of the model in the text travel scene can be improved; the semantic analysis and multi-round guiding capability of the model on fuzzy input is improved; and new and old knowledge fusion is balanced.

The invention belongs to the field of computer vision and artificial intelligence security, and discloses a self-adaptive frequency domain adversarial training method and device for a target detector, the target detector comprises a repair module and a pedestrian detector which are connected in series, and the input of the repair module is connected with the output of a patch detector; the self-adaptive frequency domain adversarial training method comprises the following steps: losses in joint training comprise standard target detection losses, repair consistency losses on a frequency domain based on a frequency domain image corresponding to a training image and a clean image, and repair dependence losses based on a detected average precision mean value; according to the invention, the end-to-end joint training is carried out through the restoration module and the subsequent pedestrian detector, and the optimization target of the restoration module is directly aligned with the improvement of the detection robustness, so that the confrontation disturbance is eliminated as far as possible, and meanwhile, the key semantic information of the detection task is reserved to the maximum extent. The separation of the performance of the repair module and the pedestrian detector is avoided, and the detection robustness is improved.