6results about How to "Shorten analysis time" patented technology

This invention relates to the field of chromosome abnormality recognition technology, specifically to a chromosome abnormality detection method based on a multimodal large model. The method includes: constructing an image dataset and a text dataset; fusing image features and text features to obtain multimodal fusion features; assigning anomaly scores to image blocks belonging to band regions based on anomaly scoring rules formulated from the multimodal fusion features, and comprehensively processing the anomaly scores of all image blocks corresponding to the chromosome to determine whether the chromosome image is abnormal; and decoding and generating natural language text that meets the requirements of chromosome abnormality detection based on the multimodal fusion feature representation and anomaly scoring rules. This invention achieves accurate chromosome abnormality detection and band location positioning through multimodal fusion and dynamic text generation mechanisms, generating interpretable natural language text descriptions, and improving the practicality and interpretability of the detection results.

This invention belongs to the field of high-dimensional vector retrieval technology, specifically involving a method for constructing and retrieving high-dimensional vector indexes based on hierarchical heterogeneous storage. The method includes an index construction stage and a query retrieval stage. The index construction stage includes hierarchical configuration of the Hierarchical Navigation Small World Graph (HNSW), hierarchical heterogeneous storage, and neighbor rule construction. The query retrieval stage includes starting node selection, PQ algorithm candidate set pre-screening, LRU cache optimization, and binary serialization storage. The method is structured as "basic storage - pre-screening - intermediate acceleration - low-level parsing." Through the collaborative design of "hierarchical heterogeneous storage + multiple optimization mechanisms," this method significantly improves query throughput during retrieval while controlling memory overhead. It is suitable for AI applications in high-dimensional vector scenarios such as retrieval enhancement generation, personalized recommendation systems, and multimodal search, where there are strict requirements for memory usage, query throughput, retrieval recall, and dynamic update efficiency.

This invention discloses a method for determining the elemental content of iron ore using variable importance-artificial neural network-assisted laser-induced breakdown spectroscopy. The method includes the following steps: S1: Collect LIBS spectral data and elemental content data from at least 13 batches of iron ore in at least four categories. Use radiometric analysis (RF) to measure the importance of LIBS spectral features. A variable importance threshold is used to optimize the input variables of the BP-ANN model. The variable importance threshold and the number of neurons in the BP-ANN model are optimized using the determination coefficients and root mean square error of five-fold cross-validation to establish a VI-BP-ANN model. S2: Using the VI-BP-ANN model from S1, input the LIBS spectral data. The model ranks the features according to variable importance and calculates the elemental content in the iron ore. This invention's detection method can rapidly detect the total iron content, calcium content, magnesium content, aluminum content, and silicon content in iron ore.

The application discloses a kind of caisson construction dynamic monitoring and deviation correction linkage method, comprising: in caisson structure and peripheral environment preset multiple type monitoring point, real-time acquisition monitoring data;Based on real-time monitoring data, when caisson inclination, axis displacement, daily sinking volume any one parameter reaches design allowed threshold value, deviation correction start signal is automatically triggered;According to deviation cause, deviation actual value and design allowed threshold value, corresponding deviation correction measure combination is matched;Deviation correction measure implementation process continuously acquires monitoring data, real-time verifies deviation correction effect, until caisson posture recovers to within allowed deviation range;Until caisson elevation, inclination, axis displacement simultaneously satisfy design requirement, stop deviation correction, enter bottom sealing procedure.The application also discloses a kind of caisson construction dynamic monitoring and deviation correction linkage system.The application is by real-time monitoring caisson posture, structure stress, peripheral stratum deformation, combined with symmetrical excavation, local counterweight deviation correction, realizes sinking precision dynamic regulation and control.

The invention relates to a vehicle software flashing effect evaluation method and device based on association attribution analysis. The method comprises the following steps: acquiring multidimensional data flashed by vehicle software in real time; through a Pearson correlation coefficient and a Granger causal test, an incidence relation matrix is generated; clustering the multi-dimensional data through a K-means clustering algorithm to obtain key parameters of the multi-dimensional data; on the basis of the incidence relation matrix, the key parameters and the flashing operation, through a Binseg algorithm and geometric topology characteristics, constructing a space-time evidence chain of vehicle software flashing; based on the space-time evidence chain, through Shapley value calculation, determining a plurality of flash factors; eliminating interference factors from the plurality of flashing factors through matching contrast and logistic regression of multi-dimensional data; and on the basis of the plurality of flashing factors after the interference factors are eliminated, calculating a software flashing score through an analytic hierarchy process. According to the method, through multi-dimensional data association analysis, evidence weight optimization and contribution degree quantification, the evaluation accuracy, the operation and maintenance efficiency and the flashing success rate are improved.

The invention discloses a fault analysis method and system based on a virtual DCS, the method is applied to a fault analysis system, and the fault analysis system comprises an injection device, a deduction analysis device and a nuclear power simulation platform based on the virtual DCS. The DCS fault data are sent to the injection equipment, the DCS fault data comprise fault points, monitoring variables and deduction time, the injection equipment generates fault variables, control instructions and monitoring instructions according to the fault points and the deduction time, and the fault variables and the control instructions are sent to the nuclear power simulation platform; the monitoring variable and the monitoring instruction are sent to the deduction analysis equipment, the nuclear power simulation platform executes simulation according to the fault variable and the control instruction, and the deduction analysis equipment monitors the simulation process of the nuclear power simulation platform according to the monitoring variable and the monitoring instruction to obtain a monitoring result corresponding to the monitoring variable. The efficiency of nuclear power plant fault analysis and the accuracy of a fault analysis result are improved.