329 results about "Quality assessment" patented technology

The invention discloses a casting surface treatment defect detection and quality evaluation method and system, and relates to the technical field of casting quality evaluation, and the method comprises the steps: collecting the surface data of a to-be-detected casting, and carrying out the data preprocessing, and obtaining a standardized input data set and a standardized data subset; calling a corresponding analysis sub-model for each subset, outputting quality features and confidence coefficients, and summarizing the quality features and the confidence coefficients into a sub-source result; environment state information is acquired to determine sub-model weights, and weighted fusion is carried out on sub-source results to obtain an evaluation result and an overall confidence coefficient; calculating space / feature / time consistency and judging according to a combination rule; when a re-checking condition is met, obtaining a re-checking label backflow updating data set, and training an updating model and parameters; and dynamically adjusting the threshold value and the weight value according to the performance index for subsequent evaluation. According to the method, multi-source data and environment information can be fused, weight self-adaption and closed-loop updating are parallel, accuracy and stability are improved, misjudgment and missed judgment are reduced, and complex working condition adaptability and long-term reliability are enhanced.

The invention discloses an academic paper review service management system based on an AIGC technology, and belongs to the technical field of paper review. The system comprises a paper import module, a format specification detection module, a content quality evaluation module, an academic atlas analysis module, an AIGC content detection module and a collaborative review module, and through a multi-dimensional evaluation model trained by historical review data, quantitative analysis can be performed on paper quality from core dimensions such as research design rationality and achievement innovation; an academic database is constructed in combination with a knowledge graph technology, so that transverse technical route comparison and longitudinal academic venation tracking of core arguments of papers can be realized, and innovation points and research gaps can be accurately positioned; a double-layer detection mechanism fuses traditional duplicate checking and AI generation content recognition, and the recognition problem of the AIGC text is effectively solved, so that a complete system covering paper quality evaluation, innovation point mining and academic integrity detection is formed, and the review depth and scientificity are comprehensively improved.

The invention discloses an AI-based prefabricated member multi-modal visual quality intelligent detection system and method, and the system comprises a multi-modal visual data collection unit which is used for obtaining the multi-modal visual data of a prefabricated member detection region, calibrating the multi-modal visual data, and constructing a multi-modal data set. And the feature fusion and candidate region extraction unit is used for inputting the multi-modal data set into a multi-modal feature fusion network to extract multi-modal features, fusing the multi-modal features and generating a defect candidate region. And the defect type identification and quantitative analysis unit is used for calling a deep learning detection and segmentation model to carry out defect type classification and defect boundary segmentation on the defect candidate region, and calculating a quantitative index of the defect region in the segmentation boundary by using the three-dimensional point cloud and the infrared image. And the defect grade judgment and component quality evaluation unit is used for carrying out dynamic threshold judgment on the defect area according to the defect type and the quantitative index of the defect area, and outputting the defect severity grade and the quality evaluation result of the prefabricated component.

The invention provides a multi-dimensional quality evaluation method, device and equipment for carbon footprint data and a storage medium, and the method comprises the steps: obtaining multi-source heterogeneous data of a product, the multi-source heterogeneous data comprising basic information of the product and carbon footprint related data of the whole life cycle of the product; generating carbon footprint data of the product based on the multi-source heterogeneous data; and performing multi-dimensional quality evaluation on the carbon footprint data based on a pre-constructed carbon footprint knowledge graph, a time dimension detection model, a space dimension detection model and a cross-dimension detection rule. According to the invention, the carbon footprint knowledge graph, the time dimension detection model, the space dimension detection model and the cross-dimension detection rule are introduced to realize semantic, time, space and cross-dimension multi-dimensional quality evaluation of the carbon footprint data, and the detection dimension can cover the multi-dimensional features required by the carbon footprint calculation. Therefore, it is ensured that the carbon footprint data quality evaluation result can reflect the influence on the carbon footprint calculation reliability more accurately.

The invention relates to the technical field of aircraft manufacturing, and discloses an aircraft part production quality optimization method based on data analysis. The method comprises the following steps: collecting multi-process real-time processing parameters and quality inspection data of a production line, and generating a dynamic quality weight matrix according to a process parameter coupling degree; extracting process path difference characteristics of qualified products and unqualified products in historical batches, and encoding the process path difference characteristics as a quality evolution chain matched with adjacent process parameter mutation relevance; optimizing process parameters at a quality evaluation node, driving a parameter combination to iterate to minimize quality fluctuation, generating an adjustment amount, updating an evolution chain constraint coefficient, and synchronously constructing a stability evaluation function of a correlation weight matrix and a defect propagation path; triggering process compensation according to an adjustment amount gradient, and verifying the cohesion through a tolerance rule by taking a reference parameter matched with the target evolution chain as compensation data; and generating a feedback matrix by using the quality fluctuation index and the compensation result, and correcting the mapping relation between the weight matrix and the evolution chain.

The invention discloses a resting electroencephalogram quality evaluation method and system based on double-branch comparative learning, and the method comprises the steps: collecting an original EEG signal X, carrying out the data preprocessing and data enhancement, generating two different enhanced views, transmitting the two different enhanced views to a double-branch encoder in parallel, respectively extracting a time domain waveform and a time-frequency domain rhythm feature, and carrying out the deep fusion, the fused features X1 and X2 are sent to a projection head, and through a self-supervised contrast learning mechanism, the network weight is optimized by using contrast loss; a small amount of labeled fine-tuning data sets including X and corresponding labels y are adopted, after flowing through a pre-trained double-branch encoder, the fine-tuning data sets are directly sent to a classification head connected with the back of the double-branch encoder so as to output a prediction result of an input data segment, and a mixed loss function including classification loss and comparison loss is adopted for training in the optimization process. According to the invention, a complete online real-time quality control system is established, and end-to-end real-time closed loop from data acquisition to quality evaluation is realized.

The invention provides a data quality evaluation method and system, and relates to the technical field of data quality evaluation.The data quality evaluation method is used for coupling type data quality evaluation of knowledge graph construction and specifically comprises the following steps that an entity historical version set of a knowledge graph and attribute time sequence data of the entity historical version set are obtained, calculating a redundancy index and a timeliness entropy change index of each entity, and calculating a coupling quality index based on the coupling coefficient; constructing a global quality evaluation matrix; and outputting a quality evaluation result and a cleaning strategy according to the distribution characteristics of the singular values of the global quality evaluation matrix. According to the method, the accuracy, adaptability and operability of knowledge graph data quality evaluation are remarkably improved, and an effective technical means is provided for constructing and maintaining a high-quality, low-redundancy and high-timeliness knowledge graph.

The invention belongs to the technical field of big data, and discloses a drawing quality evaluation management system and method based on big data. Comprising the following steps: extracting meta-information of a drawing, and extracting a category identifier by utilizing file name data and professional data in the meta-information to obtain a category of the drawing; carrying out marking; key information of the drawing is obtained, and a designer risk factor, a standard conformity risk factor and a version difference risk factor are calculated; calculating an integrity evaluation value of the drawing; dividing the drawings into defect drawings and standard drawings by utilizing the integrity evaluation values of the drawings; dividing a defect drawing into a first-level defect and a second-level defect; for a first-level defect, starting a multi-dimensional feature missing rate quality evaluation method to calculate a quality score of the defect drawing; for the standard drawing, calculating a quality score of the standard drawing according to the basic quality score and the proportion matching degree; and accurate data support is provided for drawing design and examination.

A method for identifying and calculating the volume of over- and under-excavation in tunnels based on laser point clouds belongs to the field of intelligent tunnel construction and blasting quality assessment technology. Technical solution: The scanned point cloud data is denoised; a design contour point cloud is generated and aligned with the centerline of the actual excavated tunnel; the actual collected tunnel excavation point cloud and the design contour point cloud are Delaunay tetrahedralized; the actual tunnel point cloud is traversed, and the vector dot product formed between the point cloud and the triangular sides of the convex hull is calculated; over- and under-excavation areas are identified; a polyhedron is constructed from the design contour point cloud and the actual excavated area point cloud constituting the area; the polyhedron is divided into quadrilaterals, and the volume of each quadrilateral is calculated, thereby calculating the volume of each over- and under-excavation area. Beneficial effects: This invention can quickly measure over- and under-excavation areas and quantities that are difficult to calculate in tunnel blasting quality assessment, and can replace cumbersome and time-consuming total station and manual measurement methods; this method has the advantages of high precision, high efficiency, and high flexibility.

The invention provides a data analysis method and system based on case self-learning. The method comprises the steps of obtaining a target task input by a user; retrieving a historical case based on the target task; if the historical case is retrieved, extracting context parameters according to the target task to adapt a solution of the historical case, generating a customized solution and obtaining a first analysis result; if not, performing real-time planning through a large language model to generate a new solution and obtain a second analysis result; performing quality evaluation on the newly generated solution and the second analysis result; and when the evaluation result meets a preset storage condition, storing the new solution as a new historical case in the case library. Through double-path processing and quality evaluation feedback closed loop, the problems that in the prior art, a case library is prone to being polluted due to lack of quality evaluation, and multiplexing of historical cases is rigid and low in efficiency are solved, the self-learning quality can be guaranteed, and the analysis efficiency can be improved.

The invention provides an experience knowledge graph construction and question answering method based on heuristic self-question answering, and the method comprises the steps: randomly dividing a text corpus into small batches, inputting the small batches into a large language model, and obtaining a heuristic rule set through induction processing and quality evaluation; constructing question and answer pairs by using the heuristic rule set and the text content cues, converting the question and answer pairs, and then carrying out confidence calculation to obtain a structured experience triple with confidence; constructing a knowledge graph by using the structured experience triad with confidence; extracting the core fragment of the original narrative query by using the heuristic rule set and generating a refined query text; obtaining an empirical path by using the embedded vector of the refined query text and the knowledge graph; and inputting the original narrative query and the experience path into the large language model to generate a final answer. According to the method, logic confusion and content contradiction possibly caused by a traditional RAG method are effectively avoided, the generated answer is more reliable, and the reasoning process is more interpretable.

The invention discloses a DeepSeek-based role large model fine tuning corpus automatic generation method. The method comprises the steps of receiving a role setting document describing target role features and an optional domain knowledge base; utilizing a Prompt engineering module to construct a generation Prompt containing deep role injection information and context perception knowledge fusion based on a role setting document and an optional domain knowledge base; the deep role injection information is used for enabling the generated content to accord with characters, mood, styles, knowledge backgrounds and value views defined in the role setting document; calling a preset basic large model core, and performing text generation by using the Prompt to obtain candidate corpora; performing automatic multi-dimensional quality evaluation on the candidate corpora by utilizing a quality evaluation module; wherein the assessment dimensions of the multi-dimensional quality assessment include role consistency, fact accuracy, diversity and security; and screening out qualified corpora based on an evaluation result of the multi-dimensional quality evaluation and a preset strategy, and forming a final high-quality corpus. The invention further discloses a system, electronic equipment and a computer readable storage medium.

The invention discloses a quality evaluation method and system for a composite board, and relates to the technical field of quality detection, and the method comprises the steps: constructing a latent variable feature space representing interface binding energy abnormity based on an initial response data field through a physical constraint relation between interface local response features and binding energy distribution; extracting a topological correlation structure between local latent variable features in the latent variable feature space through a graph convolutional network to obtain a local binding energy anomaly initial candidate region; active response mode enhanced excitation is carried out on the local area of the initial candidate area, the dynamic enhanced response of each local area is obtained, and a binding energy abnormal area is determined; determining the specific position and range of the hidden pseudo solder strip according to the correlation topological structure difference between the local feature in the binding energy abnormal region and the normal region; according to the method, precise identification and high-precision positioning of the titanium-steel composite plate interface hidden cold solder strip are realized.

The invention discloses a comprehensive quality evaluation method and system for Beidou deformation monitoring observation data in a complex environment, and relates to the technical field of satellite navigation positioning, and the method comprises the steps: carrying out the comprehensive evaluation of an elevating angle, a signal-to-noise ratio, an epoch difference and an environment model of the observation data of a Beidou single station, and determining the initial quality of the data; based on the double-difference observation values among the multiple stations, the data quality evaluation precision is further improved by combining a double-difference ambiguity resolving method; according to the method, ionosphere delay, three-difference errors and inter-epoch difference errors are used for data fusion, a more accurate quality evaluation result is obtained, and finally, the data are optimized through comprehensive analysis based on the quality evaluation results of various types of data so as to ensure accurate positioning and deformation monitoring. The technology aims at improving the accuracy and reliability of data quality evaluation by comprehensively analyzing influence factors of Beidou observation data in a complex environment, and provides support for geological disaster monitoring, smart city construction and other high-precision application scenes.

The invention discloses a model-based sewage treatment plant data quality control and process evaluation analysis method and system, and belongs to the technical field of sewage treatment. The method aims at solving the problems that a current sewage treatment plant data acquisition system faces data missing, abnormity and deviation, and data quality control and process evaluation are separated from each other, the intelligent degree is low, and the adaptability is poor in the prior art. The method comprises the following steps: firstly, comprehensively evaluating original data by using a data quality evaluation model, and accurately identifying and processing various problem data; a complex association relationship among process parameters is deeply mined through a process association analysis model, and a whole-flow process knowledge graph is constructed; and finally, in combination with a data quality evaluation result and the process knowledge graph, realizing accurate evaluation and optimization adjustment of the process operation state. The data quality management and control level of the sewage treatment plant is effectively improved, the accuracy of process evaluation is enhanced, powerful support is provided for stable operation and efficient management of the sewage treatment plant, and the sewage treatment industry is promoted to develop towards the intelligent and refined direction.

The embodiment of the invention discloses a test question output method and device based on multi-model collaboration, electronic equipment and a medium. A specific embodiment of the method comprises the steps of performing multi-source data fusion on a multi-source test question association data set, and then performing knowledge point association analysis to obtain a dynamic test question knowledge point graph; performing multi-hop graph reasoning query on the dynamic test question knowledge point graph to obtain a test question associated knowledge point set; performing model collaborative generation on the test question generation request information and the test question associated knowledge point set to generate an initial test question set; performing quality evaluation on the initial test question set to obtain a test question quality evaluation value set; dynamically adjusting the initial test question set to obtain an adjusted test question set and storing the adjusted test question set; and carrying out expansion adaptive test paper group rendering display on the adjusted test question set to obtain a test question test paper and printing the test question test paper. According to the embodiment, the test question quality can be improved, the test question generation efficiency is improved, the test question generation duration and display duration are shortened, the test question display quality is improved, and waste of display resources is reduced.

The invention discloses a software quality evaluation method based on dependency chain quality conduction, relates to the technical field of software quality evaluation, and is used for solving the problem that quality conduction between software is difficult to quantify. According to the method, the dependency chain graph covering the component dependency relationship, the version pedigree and the quality index is constructed, the evidence reliability label and the license list are combined, the target component and the conduction path which do not meet the quality evaluation threshold value are accurately recognized, the semantic mapping of the target component and the business function domain is constructed, and in the path conduction analysis stage, the quality of the target component is analyzed. A constraint propagation and causal checking mechanism is introduced, influence is transmitted layer by layer, a real path and an interference path are automatically identified, a comprehensive quality measurement item and a quality grade label of a target component are generated, when evaluation is not met, an upgrading, replacing, isolating or rollback scheme is automatically constructed, compatibility, time window and test influence of the scheme are evaluated, and a test result is obtained. And a closed-loop repair path and a treatment guide are output, so that the response efficiency and decision-making reasonability of the software quality problem are improved.

The invention discloses an interpretable multi-dimensional CI index dynamic scoring and risk assessment method, and belongs to the technical field of compliance risk assessment. According to the method, firstly, multi-source heterogeneous compliance data are collected, cleaned, subjected to caliber alignment and subjected to quality evaluation, and standardized data are obtained; a hierarchical index semantic system is constructed based on the standardized data, and a core index set is obtained through multi-dimensional screening; then, constructing a multi-dimensional dynamic threshold, calculating a weighted score in combination with the core index set, quantifying a score confidence interval through a Bayesian confidence quantification model, and mapping a risk level to form complete score data; and finally, carrying out interpretability analysis on the complete scoring data, and generating hierarchical interpretation texts and visual display contents. According to the method, systematization, precision and transparency of compliance risk assessment are realized, the suitability and credibility of a scoring result are improved, and reliable support is provided for compliance decision making.

The invention discloses a thermal power plant multi-source heterogeneous data quality evaluation method and system, and relates to the technical field of data evaluation, and the method comprises the steps: carrying out the comparison based on an alignment path length, a preset path length threshold value and a deviation threshold value, obtaining a consistent data flow and a conflict data flow, the method comprises the following steps: respectively calculating an accuracy index, an integrity index and a timeliness index for consistent data flow and equipment log data, obtaining a three-dimensional quality index matrix, weighting the three-dimensional quality index matrix through an entropy weight method, and calculating a comprehensive quality score of the multi-source heterogeneous data in combination with a real-time weight value of the multi-source heterogeneous data, and comparing the comprehensive quality score of the multi-source heterogeneous data with a preset score threshold to obtain a quality evaluation result, and directly generating a conflict alarm signal for a conflict data flow. According to the evaluation method based on the data indexes, not only is the sensitivity to data logic contradictions enhanced, but also context awareness and weight optimization in the evaluation process are realized.

The invention provides a geoscience specimen digital acquisition and management system, and relates to the technical field of geoscience specimen acquisition, and the system comprises a data acquisition module which collects specimen attribute fields, specimen photos and process metadata; the data quality evaluation module is used for performing bidirectional modulation on the basic weight based on the dependency relationship set of the fields to obtain an effective weight, and calculating an integrity score based on the effective weight; calculating an accuracy score, a specification degree score and a timeliness degree score; calculating to obtain a comprehensive quality score; the data optimization suggestion module is used for generating an optimization suggestion according to the score of each dimension and the score of the comprehensive quality; and the data storage management module is used for storing specimen attribute fields, specimen photos, process metadata, quality evaluation results and optimization suggestions. According to the method, the quality problem can be recognized in time, targeted optimization suggestions can be automatically generated, and intelligent priority recommendation is provided for field acquisition based on historical quality statistical data, so that the quality and efficiency of geoscience specimen data acquisition are improved.

The invention relates to the technical field of giant salamander by-product quality evaluation, and discloses a giant salamander by-product source functional factor quality evaluation system and method, and the method comprises the steps: obtaining a giant salamander by-product sample, carrying out the pretreatment, and extracting a functional factor to obtain a to-be-detected sample; identifying and quantitatively determining specific markers based on biochemical analysis and molecular detection means to generate a marker data set; inputting the marker data into a multi-modal feature fusion module, extracting activity and stability features, and performing time sequence dynamic modeling; activity and stability independent evaluation branches are set, and corresponding score values are generated; dividing functional domains according to functional factor categories, constructing an evaluation curve based on historical data, substituting a target sample into each functional domain and an adjacent functional domain to generate functional domain evaluation results, and performing weighted fusion; and finally, carrying out weighted integration on the activity score, the stability score and the comprehensive evaluation value, and determining a final quality evaluation result.

The invention relates to the technical field of data processing, and discloses a road inspection data quality real-time monitoring and abnormal data labeling method and system, and the method comprises the steps: carrying out the multi-dimensional cleaning of the road inspection data, and obtaining the standard data of the road inspection data; comprehensively judging the real-time state of the standard data based on a preset data quality dimension to obtain the integrity state and the accuracy state of the standard data; performing quantitative evaluation on the integrity state and the accuracy state based on a quality rule base of the road inspection data to obtain a quality evaluation result of the road inspection data; performing multi-dimensional comparison on the quality evaluation result and an abnormal data segment in a historical dynamic threshold value to obtain an abnormal identifier of the road inspection data; according to the abnormal level of the historical road inspection data, mapping the abnormal identifier to standard data to obtain structured abnormal annotation data of the road inspection data; according to the invention, the efficiency of real-time quality monitoring and abnormal data labeling of the road inspection data can be improved.

The present invention discloses a tightening quality assessment method based on deep learning, which relates to the technical field of tightening quality assessment. The present invention collects torque, preload, lubrication status and environmental data in real time during the tightening process, uses an LSTM dynamic model to capture the nonlinear fluctuation of the friction factor during repeated tightening, and promptly corrects the torque-preload conversion relationship. Adaptive learning rate and fully connected layer mapping are adopted to enable the model to automatically update parameters under different environmental conditions, dynamically adjust the tightening speed, preset torque and lubrication strategy. The environmental integration module quantifies changes in temperature, humidity and salt concentration, corrects process parameters in real time, and effectively suppresses instantaneous abnormal fluctuations.

The invention discloses a dam continuous compaction monitoring method and system, and relates to the technical field of dam rolling monitoring, and the method comprises the steps: installing a plurality of sensors on dam rolling equipment, and collecting dam rolling data in real time; meanwhile, environment data of a dam filling area are collected; constructing a data fusion model, performing fusion analysis on the dam rolling data by adopting a data fusion algorithm, and extracting characteristics related to the compaction quality; correcting the extracted features through environment data by adopting an environment disturbance algorithm to obtain monitoring data; and based on a preset compaction quality standard and a threshold value, evaluating the monitoring data to obtain a compaction quality evaluation result, and feeding back the compaction quality evaluation result to an operator. Real-time, continuous and accurate monitoring of the dam rolling process is achieved, and the quality problems of excessive rolling, insufficient rolling and the like are found and solved in time.

The invention relates to the technical field of data quality evaluation, in particular to an automatic quality evaluation method for a high-quality data set in the cultural field, which comprises the following steps: extracting semantic fragments and vector information, performing clustering analysis on a coverage structure, screening redundant samples, labeling image-text differences, and outputting a consistency evaluation result. According to the method, high-precision recognition of semantic coverage and structure density can be realized through structured extraction of semantic information and vector distribution measurement, and the discriminability of redundancy detection is enhanced through linkage screening of vector coincidence frequency and similarity threshold. The systematicness of image-text consistency marking is improved through an image-text position difference and semantic span combined marking mechanism, and it is ensured that redundant and deviated intersection samples can be accurately recognized through a synchronous contrast mechanism between text screening content and image-text difference marking results. The intensive clustering of the cultural semantic content in the vector space and the quantitative analysis capability of the structure deviation point are synchronously improved, and the comprehensive dimension and logic closure of the evaluation result are optimized.

The invention relates to the technical field of AI reasoning optimization, and discloses a user AI reasoning service method based on data quality intelligent evaluation, which comprises the following steps: a target AI model receives reasoning request data input by a user and executes data format and integrity verification, and the data comprises text, image or sensor time sequence data; performing multi-dimensional quality evaluation on the inference request data passing the verification, and aggregating and generating a data quality score for judging data reliability by extracting quality indexes and analyzing data features; the reasoning precision of a target AI model is dynamically regulated and controlled in real time based on the data quality score, so that prediction deviation caused by input data noise is inhibited; and executing an inference task by using the regulated and controlled target AI model, generating a prediction result, adding quality evaluation metadata, and returning the prediction result to the user terminal. According to the method, the reasoning precision of the AI model is regulated and controlled in real time through the data quality score, the problem of prediction deviation caused by input data noise is solved, and the robustness of the model under different quality data is improved.

The invention provides a test data quality evaluation method and system based on machine learning, relates to the technical field of data processing, and solves the technical problems of low data quality evaluation efficiency, non-systematic evaluation dimension and lagging abnormal response of vehicle test data in the prior art. The method comprises the following steps: acquiring test data of test equipment through a data acquisition interface; the test data comprises real-time test data and historical test data; performing quality evaluation on the test data based on a preset evaluation dimension by using the trained data quality evaluation model; the evaluation dimensions at least comprise data integrity, data accuracy, data consistency and data validity; and generating a data quality evaluation report according to an evaluation result, and performing exception marking and early warning pushing on the data which is evaluated to be unqualified.

The invention provides a CORS network data quality evaluation method based on entropy weight TOPSIS. The method comprises the following steps: S-1, forming an original matrix; s-2, standardizing the xij to obtain a standardized Zij (Zij); s-3, calculating the proportion of the index j corresponding to the object i in the column sum; s-4, calculating an entropy value of an index j corresponding to the object i; s-5, calculating a difference coefficient of the jth index; s-6, calculating the weight of the jth index; s-7, multiplying the weight point calculated in the step S6 by the standardized matrix to obtain a weighted standardized matrix; s-8, positive and negative ideal solutions are determined; s-9, calculating an Euclidean distance; s-10, finally calculating a comprehensive score Si of the base station in the ith CORS network; and S-11, judging the final advantages and disadvantages of all the reference stations of the current CORS station network according to the score. The GNSS and CORS data quality evaluation method combines observation data acquired by a CORS station network in real time or after the event and entropy weight TOPSIS with objectivity, comprehensiveness and high efficiency as core advantages, and provides quality evaluation for GNSS and CORS data.

The invention relates to the technical field of artificial intelligence, and discloses a training data generation method and device, computer equipment and a storage medium, and the method comprises the steps: obtaining context information of a question and answer dialogue and original reply data; taking context information of the question and answer dialogue as input, taking the original reply data as a cue word, and based on a pre-constructed first open source model, generating thinking process data; based on a pre-constructed second open source model, performing quality evaluation on the thinking process data to obtain an evaluation result; and when the evaluation result passes, adding the thinking process data into a training data set, and when the evaluation result does not pass, discarding the thinking process data. The method can be applied to business platforms of financial science and technology, medical health and the like, and solves the technical problems of style distortion of question and answer models distilled by a distillation technology and lack of a quality evaluation system in a thinking process in the prior art.

The application discloses a kind of distillation algorithm multi-modal data quality evaluation method.The present application is by including data quality adjustment coefficient in model training, the sensitivity of model degree data quality, data compliance is strengthened, the training efficiency of model is improved, the professionalism of model is enhanced;The training results of student model are fed back to teacher model through knowledge reflux, so that the teacher model can be better optimized, and the new student model is distilled through the optimized teacher model, forming a positive cycle mechanism of coevolution, so that the teacher model and student model can be continuously optimized, and the accuracy of future data quality evaluation is improved;And student model is obtained by distillation, and is actually deployed to customer data end for data quality evaluation, which consumes limited resources, and as localized deployment, data does not need to be transmitted through the Internet, solving the problem of excessive data transmission and low data quality evaluation efficiency.