425 results about "Data synthesis" patented technology

The invention provides a training data synthesis method and device based on error extrapolation and inference chain analysis, a medium and a program product. The method comprises the steps of obtaining an initial sample set; performing multiple sampling reasoning on the problem of each task sample by using a small language model to generate a plurality of reasoning chains; calculating an overall error score of each reasoning chain based on a preset error evaluation rule, and determining a to-be-corrected reasoning chain; the inference chain to be corrected and the corresponding question are input into the large language model together, and a corrected answer is generated; forming a new task sample by the question and the corrected answer, and finely adjusting partial parameters of the small language model; repeatedly executing the process until the performance index change rate of the model on the task evaluation set is lower than a preset threshold value, and outputting a final task sample; and forming a training sample set by a plurality of final task samples, and performing all-parameter fine tuning on the small language model. According to the method, the training data self-optimization path is constructed by taking the model error as guidance, so that the semantic consistency and the data validity are improved.

The invention relates to the technical field of self-adaptive control, in particular to a tobacco field variable fertilization control system and method, and the system comprises a sensing data collection module which collects soil humidity data through a soil humidity sensor, monitors the growth data of crops at the same time, and carries out the synthesis to obtain real-time soil and growth data; and performing time sequence analysis on the real-time soil and growth data, identifying a data trend and an abnormal point, and generating a trend analysis result. According to the method, the soil humidity and the crop growth data are collected, and the time sequence analysis is combined, so that the trend and the abnormality are quickly identified, and the prediction accuracy is improved. The convolutional neural network is used for carrying out recognition and probability evaluation on diseases and insect pests, so that disease and insect pest management is more accurate, and response is quicker. Through the technologies, the system can dynamically adjust irrigation parameters based on real-time data, optimize water distribution, and ensure that an irrigation strategy is perfectly matched with crop requirements and potential risks.

Techniques for determining a presence of an object, especially an object such as animal or debris, in a path of a vehicle, are discussed herein. For example, sensors of various modalities, which may include multispectral sensors, may capture data representing an environment the vehicle is traversing. In examples, one or more trained machine learned (ML) models, operating on a vehicle computing system, may detect and / or classify objects in the environment, based on input data of one or more modalities or spectral bands. The ML models may be pre-trained using training data including real sensor data, synthetic data, and / or augmented data, along with auto-generated annotations. In some examples, hyperspectral data may be used to identify materials associated with detected objects. A confidence score associated with the detection of the object may also be computed. The vehicle may be controlled based on detection of the object and its classification.

The invention relates to a reasoning data synthesis method and system, and the method comprises the steps: obtaining a mathematical big language model, and carrying out the question fine adjustment of the mathematical big language model through a question fine adjustment method by employing a preset data set in a model training process, and obtaining the mathematical big language model processed through the question fine adjustment method; generating a sample problem through the mathematical big language model processed by the problem fine tuning method; the sample problem is optimized through a problem preference optimization method, and then the optimized sample problem is used for training the mathematical big language model processed through the problem fine adjustment method; taking the mathematical big language model processed by the problem fine tuning method and the problem preference optimization method as a problem generator, and filtering problems generated by the problem generator; a corresponding answer is generated for each filtered question to synthesize inference data. The reasoning data synthesis method not only can ensure the data quality, but also can reduce the generation cost and improve the generation efficiency.

The invention provides a three-dimensional data synthesis method, electronic equipment, a storage medium and a program product, and the method comprises the steps: obtaining a plurality of texture-free three-dimensional grid models, carrying out the rendering of each three-dimensional grid model, obtaining a plurality of view images, and constructing a training sample pair in combination with a semantic description text; based on the training sample pair, utilizing a low-rank adaptation technology to carry out fine tuning on the first text graph model to obtain a second text graph model with multi-view consistency understanding ability; modeling a latent space semantic difference between the rendered image of the to-be-deformed three-dimensional grid model and the target semantic text based on the second text graph model, constructing an optimization constraint and updating parameters, and obtaining a deformed three-dimensional grid model conforming to target semantics; and generating a texture image by using the texture generation model, and attaching the texture image to the surface of the deformed three-dimensional grid model to obtain synthetic three-dimensional data with textures. The reconstruction precision, the diversity expression ability and the cross-category generalization ability of the synthesized three-dimensional data are obviously enhanced.

The invention discloses a radio interference identification method based on electromagnetic spectrum monitoring, and the method comprises the following steps: S1, collecting spectrum data in an electromagnetic environment through broadband spectrum monitoring equipment, and carrying out the digital sampling processing; s2, adopting a diffusion probability model to automatically remove environmental noise and non-interference signals; s3, automatically generating an expansion data set of edge and small sample interference signals by adopting an attention-enhanced data synthesis expansion technology; s4, extracting multi-dimensional time-frequency domain dynamic correlation characteristics of the spectrum interference signal based on a Transform structure; s5, adopting an attention enhancement self-supervision algorithm to generate a self-supervision soft label of an unknown interference type; s6, dynamic back diffusion iteration is carried out through the diffusion probability model, and an interference classification result is obtained; and S7, constructing a closed-loop dynamic feedback mechanism by using an interference classification result. According to the invention, accurate identification of radio interference signals is realized, and the accuracy, dynamic adaptability and automation degree of interference signal classification are improved.

The invention relates to the technical field of artificial intelligence, and provides a multi-modal risk data synthesis method for anti-fraud and risk identification. The method comprises the following steps: acquiring a target face image, target sound data and a video containing a face and sound; in the acquired video or image, replacing an original face with a target face through a deep counterfeiting technology, and replacing original sound in the video with target sound; marking images and audios in the generated multi-modal data; and inputting the forged and synthesized multi-modal data and the labels thereof into a multi-modal large model, identifying and analyzing the multi-modal data, and generating a risk abstract. According to the method, a systematic multi-modal risk content generation process is constructed, the requirement of a multi-modal large model for high-quality training data is met, the advantages of synthetic data in risk analysis are brought into full play through combined use of forged data and the multi-modal large model, and the risk analysis efficiency is improved. And the identification and coping capability of potential risks in a complex multi-modal scene is obviously improved.

The invention provides an anti-fact fair data synthesis method and device based on causal reasoning, and aims to generate high-quality synthesis data meeting the fairness requirement by mining the causal relationship between observable features. The synthesis method comprises the following steps: extracting observable features, sensitive features and labels from original data, extracting potential features through a variational automatic codec, and constructing a causal relationship graph; designing a generator according to a topological sequence of the causal relationship graph, connecting a causal path, inputting the potential features and the related features into the generator in sequence, and constructing a data generation process conforming to a causal structure; introducing a discriminator to carry out adversarial training on a generation result and original data, and optimizing generator parameter distribution; finally, synthetic data meeting fairness requirements are generated. According to the method, effective regulation and control on the influence of sensitive characteristics and strict constraint on a causal structure are realized, the generated data has higher fairness and interpretability, and the method can be applied to the fields with higher fairness requirements, such as finance, medical treatment and education.

The invention belongs to the technical field of nuclear radiation measurement, and relates to a liquid flash TDCR multi-nuclide beta spectrum analysis method and system based on artificial intelligence. The method comprises the following steps: constructing a numerical model of a liquid flash detector spectrometer by using a Monte Carlo technology to simulate the energy spectrum response of a single nuclide under different quenching conditions; constructing a training database by adopting a parameterized data synthesis algorithm; constructing a multi-task mixed spectrum analysis neural network model; carrying out model training by utilizing the constructed database; and inputting an actual measurement spectrogram, and outputting the multi-nuclide absolute activity, the detector efficiency and the decomposition energy spectrum contribution curve in real time by using the trained neural network model. According to the method, differential distribution characteristics and quenching response curve characteristics of the beta continuous energy spectrum are creatively fused, a neural network architecture with physical mechanism constraints is constructed, and liquid flash TDCR multi-nuclide energy spectrum characteristic decoupling and accurate activity solving are achieved under the unknown quenching condition.

The embodiment of the invention provides a data synthesis method, a model pre-training method, a data synthesis device, a computer readable storage medium and electronic equipment, and relates to the technical field of artificial intelligence. The method comprises the following steps: splitting a target task corresponding to a target pre-training model to obtain a plurality of sub-tasks; according to seed data in the seed data set, cue words are determined, the data in the seed data set are data about the target task collected from multiple channels, the cue words are cue words corresponding to the subtasks or cue words corresponding to the target task, and the cue words have multiple styles; the cue words are input into a big language model, the big language model is a trained model, and generated data output by the model correspond to the cue words; and determining synthetic data about the target task according to the generated data. According to the embodiment of the invention, large-scale and high-quality synthetic data can be provided, and the requirement of pre-training for large-scale and high-quality data can be met.

The invention discloses a medical text data synthesis method based on role driving, and relates to the technical field of text data synthesis and artificial intelligence, and the method comprises the following steps: S100, generating an original problem; s200, screening original problems; and S300, synthesizing the medical text data. According to role description and interest content generation, the adaptability to different medical role requirements is enhanced, semantic duplicate removal and correlation evaluation are realized by utilizing a text coding model and a vector database, semantic duplicate content is removed, correlation is accurately evaluated by utilizing a large language model, the data quality and diversity are improved, and the method is suitable for popularization and application. Through generation of variant questions and answers, diversity and practicability of medical text data are enriched. According to the method, the personalization and pertinence of the medical text data are improved, the quality and duplicate removal efficiency of the medical text data are enhanced, the problem types are enriched, the generated content is optimized, and various information requirements of the medical text data are comprehensively covered.

Techniques for determining a presence of an object, especially an object such as animal or debris, in a path of a vehicle, are discussed herein. For example, sensors of various modalities, which may include multispectral sensors, may capture data representing an environment the vehicle is traversing. In examples, one or more trained machine learned (ML) models, operating on a vehicle computing system, may detect and / or classify objects in the environment, based on input data of one or more modalities or spectral bands. The ML models may be pre-trained using training data including real sensor data, synthetic data, and / or augmented data, along with auto-generated annotations. In some examples, hyperspectral data may be used to identify materials associated with detected objects. A confidence score associated with the detection of the object may also be computed. The vehicle may be controlled based on detection of the object and its classification.

The invention discloses a single crystal turbine blade crystal orientation defect detection system based on adaptive optics, particularly relates to the field of nondestructive testing, and is used for solving the problem that thermal damage and signal consistency are difficult to consider in acoustic imaging of complex structural parts. Through partition energy threshold setting, dynamic pulse planning, photoacoustic processing and unified coordinate cross-subarea data synthesis, local thermal damage caused by excessive concentration of energy is effectively avoided in the material detection process, and the requirement for collecting high-bandwidth signals can be met in a large range. The signal-to-noise ratio and the detection efficiency are improved by regulating and controlling pulse energy and a scanning path, the amplitude comparability and the image coherence among regions are ensured through signal normalization correction and subregion image set splicing, high-resolution and comprehensive-coverage acoustic mapping is output, the crystal orientation defects of the single crystal turbine blade are positioned and evaluated, the detection precision is improved, the measurement period is shortened, and the method is suitable for large-scale popularization and application. And the method has an applicable value for aero-engine parts requiring rapid and high-precision nondestructive detection.

The invention discloses an electronic medical record data synthesis method based on semantic processing, and relates to the technical field of medical informatization, and the method comprises the following steps: S1, constructing a probabilistic medical knowledge graph; s2, generating a semantic representation vector; s3, constructing a multi-dimensional dynamic space-time atlas; s4, generating a discrete personalized disease course event sequence with space-time coordinates; s5, taking the discrete personalized disease course event sequence and the corresponding medical entity semantic representation vector as condition input, guiding the improved TSDiff model to execute an iterative denoising process, and outputting a multi-dimensional random disease course trajectory; s6, forming multi-modal electronic medical record data; and S7, performing multi-dimensional quality evaluation on the multi-modal electronic medical record data. According to the method, the limitations of logic inconsistency, modal splitting and model capability solidification in a traditional synthesis method are overcome, and an efficient and accurate solution is provided.

The invention discloses a training data generation method and device for a network attack recognition model and electronic equipment, and relates to the field of network security, and the method comprises the steps: collecting to-be-recognized emails from a plurality of information sources, carrying out the classification processing and labeling processing of the preprocessed to-be-recognized emails, obtaining a labeled email sample set, and storing the labeled email sample set in a database; extracting content features, sender features, structural features and attachment features to obtain a sample feature set, performing data synthesis by using a generative adversarial network model, generating a synthesized mail sample, converting the labeled mail sample and the synthesized mail sample into a preset model data format, generating a model training data set, training a preset large language model, and obtaining a new mail sample; and identifying whether the contact emails are phishing emails or not through the model. According to the method and the device, the technical problems that omissions of mail filtering are easily caused and the user satisfaction is influenced because a model for identifying phishing mails cannot update training data for identifying an attack mode in time in related technologies are solved.

The invention discloses an operation planning optimization data synthesis method and device based on high-quality seed data, and relates to the field of operation planning optimizing.The operation planning optimization data synthesis method comprises the steps that an initial task pool is constructed according to historical project experience, the initial task pool comprises a plurality of seed data samples, each data sample comprises a problem, a model and a code, and based on the problems in the data samples, the initial task pool is constructed; and expanding to a plurality of problem scenes through a few-sample prompting technology, generating corresponding complete problem samples, and based on the problem samples, combining the mathematical models in the initial task pool, and generating mathematical models corresponding to the corresponding problem samples through the few-sample prompting technology. According to the method, scene expansion of problem samples is realized, more practical application scenes and optimization requirements can be covered, one-to-one mapping of problems and mathematical models is realized, and the problem that model design depends on artificial experience in traditional operation optimization is solved.

The embodiment of the invention provides a risk zoning method and device for helicopter flight, equipment and a storage medium. The method is applied to the technical field of data processing, and comprises the following steps: carrying out classification and threshold division on meteorological elements according to a fusion data set and different flight scenes to obtain a meteorological element threshold breakpoint set; based on the meteorological element threshold breakpoint set, performing frequency statistics and data synthesis on the fused data set to obtain a multi-layer meteorological element frequency data set; obtaining an initial clustering result of each meteorological element through clustering analysis according to the meteorological element frequency data set; according to the initial clustering result, carrying out risk grade division on each meteorological element to obtain a standardized single-element risk zoning result; and based on a single-element risk zoning result, adopting a multi-element fusion method to obtain a comprehensive risk zoning map. In this way, the technical problem that in the prior art, due to the fact that multiple elements jointly participate in clustering, the risk level is difficult to judge can be solved.

The invention relates to an adaptive stacked workpiece point cloud data synthesis and instance segmentation method and device, and the method comprises the steps: collecting the point cloud data of a single workpiece, and carrying out the automatic reverse modeling through a feature preserving reconstruction algorithm, so as to obtain a CAD model of the single workpiece; establishing and training a stacked workpiece instance segmentation model comprising a backbone network module, a semantic prediction module, a central point offset prediction module, a binary clustering module, a centripetal offset module and a voting module; and completing the point cloud data and refining edge details through a point cloud data completion and detail feature preserving algorithm, and inputting the processed point cloud data into the trained instance segmentation model to realize accurate instance segmentation of the stacked workpiece point cloud data. According to the invention, through technologies of adaptive point cloud data synthesis, deep learning instance segmentation, point cloud data completion and the like, the quality and segmentation precision of the point cloud data are significantly improved, and the method is suitable for workpiece identification and operation tasks in a complex industrial environment.

The invention relates to the field, and particularly discloses an asset management coding model modeling method based on federated learning, and the method comprises the steps of heterogeneous asset coding alignment, a dynamic federated aggregation mechanism, small sample data source optimization and dynamic weight adjustment. Through a dynamic federation aggregation mechanism, resources are efficiently utilized, the real-time response capability is improved, through comparison between a historical state and a current state, short-term noise fluctuation is filtered, a model is prevented from being misled by an abnormal value, and rapid modeling is achieved for a newly-accessed data source through data synthesis and model pre-training. Quality-driven aggregation is improved, a high-weight data source is dominant in parameter updating, and the accuracy of the model is improved.

The invention discloses a target trajectory tracking method and product based on azimuth statistical analysis and prediction, and the method comprises the steps: converting a horizontal array manifold vector according to a set interval, carrying out the array signal processing of the array manifold vector at each angle according to a set scanning interval, and obtaining the beam width at the angle; performing spatial spectrum synthesis on the horizontal array receiving data according to a plane wave array signal processing method to obtain a horizontal array data synthesis azimuth spectrum; presetting an initial azimuth angle of a to-be-tracked target, and solving a target azimuth angle at the current moment in combination with a target operation attribute corresponding to the azimuth angle, an array azimuth estimation error and a beam width; within the statistical time, adopting a least square method to predict a target azimuth angle, and adopting a target azimuth angle track compensation value to compensate a target azimuth angle prediction value; the target azimuth angle at the current moment is extracted according to the target azimuth angle predicted value, and then continuous and stable tracking of the target azimuth trajectory is completed.

Methods and systems for generating and employing synthetic data are disclosed. The synthetic data is generated by defining roles for a plurality of speakers and inputting the roles to at least one Large Language Model (LLM), which in turn successively generates statements of each speaker which are responsive to generated statements for the other speaker based on the defined roles. Each successive set of statements are input to the LLM to generate additional statements of the speakers to obtain synthetic dialog data. The synthetic dialog data can be used to test and / or train neural networks as well as various platforms, including conversation analytics platforms.

The invention discloses a mineral spectral feature unmixing device using a generative adversarial network, which comprises a multi-modal data preprocessing module used for correcting and enhancing original mineral spectral data, solving the problem of small sample training and providing high-quality input for subsequent unmixing, and a generative network module used for fusing noise and geological text description, and providing high-quality input for subsequent unmixing. The dynamic adversarial training module is used for initializing end member features based on comparative learning of a mineral symbiosis sequence and improving priori cognition of the model on a mineral combination rule, and the end member reconstruction verification module is used for cyclically reconstructing a single albedo matrix by using a generator network unit and a discriminator network unit; nonlinear scattering and atmospheric noise interference are effectively eliminated through the model conversion unit and the 3D convolution kernel, the data synthesis capability of the conditional generative adversarial network is combined, the small sample training bottleneck is relieved, and the prior constraint of an end member feature combination rule is enhanced based on mineral symbiosis sequence comparative learning through a dynamic adversarial training mechanism.

The invention discloses a nursing scene data enhancement and generation method for intention recognition model training, and belongs to the technical field of artificial intelligence and smart old-age care. The method aims at solving the problems that in the prior art, high-quality nursing scene training data is deficient, and the obtaining cost is high. According to the core technical scheme, the method comprises the steps that firstly, a staring sequence and other context information (such as time, place and physiological signals) of a user are processed through a multi-modal information fusion model, and a structured initial situation vector is generated; secondly, inputting the vector into a scene generation model combined with a nursing knowledge base, and automatically generating a batch of basic nursing scene data with intention labels; key points are that a data enhancement module is introduced, and a plurality of innovative strategies such as situation element disturbance, virtual physiological data synthesis, gaze path variation and virtual scene deduction are adopted to deeply process basic data, so that the diversity and complexity of a data set are greatly enriched; and finally, combining the basic data with the enhanced data to construct a final comprehensive training data set. According to the method, large-scale and high-fidelity training data can be generated in a low-cost and high-efficiency manner, and the accuracy and robustness of the intention recognition model in a real nursing environment are remarkably improved.

The invention relates to the technical field of electric power prediction, discloses a photovoltaic power generation prediction system based on a multi-source analysis model, and aims to solve the problems that an existing photovoltaic power generation prediction technology is insufficient in individual and isomerized node prediction capability, and prediction is limited due to dependence on historical data in a newly-accessed and data-missing node scene. The system comprises a multi-dimensional node feature quantification module, a virtual historical data synthesis module, a power prediction module based on an enhanced data set, a closed-loop deviation traceability and correction module and a compensation strategy execution module oriented to a specific scene. Through adoption of the technical scheme, high-confidence prediction can be provided for blank or data missing nodes on the premise of not depending on historical data of the target node, and the precision, the coverage rate and the dynamic adaptability of distributed photovoltaic prediction are remarkably improved.

The invention relates to a method for generating audio deep learning training data based on an adversarial neural network (GAN), which is a data synthesis technology based on a generative adversarial network (GAN) and is used for generating audio training samples of scarce categories so as to effectively expand the diversity of a data set. According to the method, the generation capability of the GAN model is utilized to simulate the audio samples of the rare category, and the problem that the samples of the rare category are insufficient in a traditional data set can be solved. Through the technology, researchers can reduce the cost of manually collecting samples, the consumption of human resources is greatly reduced, and the universality of the data set in category coverage is ensured at the same time. In addition, the generated diversified samples can enhance the generalization ability of the model, so that the model is more stable when processing changeful data in the real world. According to the method, the training efficiency of the model can be effectively improved, the robustness of the model in different scenes can be improved, and the precision and practicability of an audio classification task in specific field application are further promoted.

One or more embodiments of the invention provide a multi-modal large model training method, apparatus and device, and a storage medium, and the multi-modal large model comprises a visual coding layer used for generating image features corresponding to an image, and a large language model used for generating a reply text based on the image features and a query text; the method comprises the following steps: aiming at aligning an image feature space of a visual coding layer and a text representation space of a large language model, adjusting a multi-modal large model to obtain an aligned multi-modal large model; based on pre-training samples which are constructed in a data synthesis mode and correspond to various tasks in the at least one task in the medical scene, pre-training the aligned multi-modal large model to obtain a pre-trained multi-modal large model; and performing fine tuning on the pre-trained multi-modal large model based on a fine tuning sample which is obtained through a manual labeling mode and corresponds to the target task in the medical scene to obtain the multi-modal large model used for executing the target task.

The invention relates to the technical field of multi-person posture estimation, in particular to an interactive scene data synthesis and key point visibility updating algorithm based on monocular vision, which comprises the following steps of: 1, arranging single-person picture data under monocular shooting vision, processing an input single-person picture under monocular shooting vision by using a GrondingDINO visual language open type target detection model, and obtaining a target detection result; and the Person is used as a retrieval keyword to identify a character individual in the picture. Through fine processing of the steps, especially introduction of target matching, target position adjustment and key point visibility updating methods, the quality of the synthesized image is greatly improved, the fusion degree of the target and the background is optimized through accurate matching and natural target pasting, and the image quality is improved. The space consistency and the visual naturalness of the synthetic image are enhanced, the problem of overfitting is effectively avoided through the improvement, the diversity of training data is enhanced, and therefore the generalization ability of the model is improved.

In disclosed techniques a system generates, using a generative model, current synthetic communications, including inputting conditions for the synthetic communications into the trained generative model. The system generates the trained generative model by iteratively performing multiple operations until a discriminator of the generative model determines that synthetic communications output by the generative model satisfy a difference threshold. The operations include: generating, by a generator of the generative model, based on existing communications, a training synthetic communications, determining, by the discriminator of the generative model, differences between the existing communications and the training synthetic communications, and updating the generator based on the differences. Using the current synthetic communications and the existing communications, the system trains another model to evaluate newly initiated communications. The disclosed data synthesis techniques may advantageously enable discovery of concealed patterns, which in turn improves detection of processing systems that execute models trained on the synthetic data.

The invention provides an embedded data synthesis method and device fusing retrieval and large model distillation and a medium. The method comprises the following steps of: preprocessing an unstructured document in a vertical field, and dividing the unstructured document into multi-granularity text blocks with a hierarchical association relationship; forming a context based on the combination of the multi-granularity text blocks, injecting disturbance information corresponding to the priori knowledge in the vertical field into the context, calling a generative model to generate a retrieval query according to the context, and determining a target text block corresponding to the retrieval query as an initial positive sample; false negative sample text blocks are filtered according to the incidence relation between the text blocks, and a positive sample set and a negative sample set are formed; and constructing a comparative learning training sample, and training the semantic representation model by using the comparative learning training sample to generate an embedded vector for the retrieval task. According to the method, the retrieval task construction efficiency and authenticity can be improved, the positive sample coverage integrity is improved, and the contrast learning training stability and retrieval precision are enhanced.

The invention belongs to the technical field of mathematical solving, and provides a mathematical application question solving method and system based on data synthesis, and the method comprises the steps: obtaining a to-be-solved mathematical application question; constructing a reasoning chain of the mathematical application questions based on a large language model; judging whether the constructed reasoning chain is converted into a program auxiliary chain or not according to the prompt judgment word; if conversion is needed, the large language model generates an inference chain and a program auxiliary chain, the program auxiliary chain is executed by using a code interpreter, and solving of the mathematical application question is completed; otherwise, solving the mathematical application problem directly according to the reasoning chain.