99results about How to "Precise screening" patented technology

The invention discloses an agent task scheduling planning method, and relates to the technical field of agent scheduling. The method comprises the steps of analyzing a task instruction to generate atomic tasks capable of being independently executed, constructing a subtask dependency graph, and defining association constraints between the tasks; determining the real-time resource occupancy state of the intelligent agent to obtain a resource state tensor, completing resource-task association anchoring and dependency priority ranking in combination with the sub-task dependency graph, and generating a task priority sequence with resource constraint; performing dynamic capability matching and predictive load balancing calculation on the sequence through a task-agent adaptation model, and determining a target execution agent of each atomic task; and based on the target execution agent and the subtask dependency graph, performing time sequence scheduling arrangement and conflict resolution, and generating a collaborative execution scheme. The method improves the reasonability and efficiency of agent task scheduling, reduces resource conflicts and execution timeout risks, and is suitable for agent cluster collaborative scheduling in a complex scene.

The invention discloses an order-driven mixed box stacking method, system and equipment and a medium. According to the method, a stacking process is decoupled into two associated subtasks of active warehouse-out decision and real-time stacking execution; the problems of frequent stack transfer and invalid reciprocating carrying caused by scattered goods combination in a physical layout limited scene in an existing offline stacking algorithm are effectively solved, and meanwhile, the space utilization rate bottleneck caused by lack of order global view in a traditional online stacking algorithm is also overcome; and a brand-new framework support and a global optimization decision view angle are provided for efficient disassembly operation in a limited physical space. An energy-based warehouse-out sequence optimization algorithm is introduced into a warehouse-out decision-making layer, and the algorithm deeply fuses the local fine sampling capability of Langevin dynamics and the global evolution capability of a genetic algorithm, so that the technical defects that a traditional heuristic rule is difficult to capture complex geometric features of goods and is easy to fall into local optimum are overcome; and the three-dimensional warehouse is guided to realize active and efficient replenishment scheduling.

This invention relates to the field of medical information management technology, including an AI-based endometriosis management system. The system comprises a thickness dynamic monitoring module, an inflammation level assessment module, a lesion spread analysis module, a periodic lesion tracking module, and a disease progression analysis module. In this invention, AI is used to segment ultrasound image data, enabling dynamic monitoring of lesion areas. Lesion boundaries are extracted, and changes in local inflammatory marker concentrations are calculated, making the screening of abnormal inflammatory areas more accurate and enabling early identification of lesion development. The tracking of periodic lesions, combined with analysis of lesion contour changes and area increases / decreases, enhances the quantitative control of disease progression. The future development trend of lesions, combined with the calculation of periodic inflammatory pathways, makes the prediction of disease evolution more consistent with physiological changes. Multi-dimensional lesion analysis combined with AI intelligent calculation expands the diagnosis and treatment of endometriosis from static assessment to dynamic trend prediction, improving the adaptability of personalized treatment plans.

The present application relates to the technical field of fault detection, in particular to a chemical production process abnormal condition identification method and system, comprising the following steps: collecting reactor temperature concentration and flow and aligning and denoising to generate a multi-source time sequence set, extracting spatial measurement point temperature difference and differentiating symbols to generate a temperature difference direction sequence, calculating conversion amount based on concentration and flow and calculating rate gradient and direction sequence inversion comparison to generate heat release inversion sequence, obtaining inversion time and spatial coordinates and misaligning to align to form a lag mismatch set, combining time window temperature change and mismatch frequency to input a support vector machine classification to obtain abnormal condition identification results, in the present application, a multi-source synchronous time sequence set is constructed to extract temperature difference direction, spatial gradient is calculated to obtain inversion characteristics, time and space misalignment is implemented to align to capture physical conduction lag law, mismatch frequency is used to rely on a support vector machine to perform classification, single point limitations in dealing with conduction lag are overcome, the technical problem that hidden faults are easily covered is solved, and accurate identification of abnormal variations in complex conditions is realized.

This invention relates to the field of nuclear radiation detection technology, and more particularly to an energy spectrum processing method for a quasi-hemispherical CdZnTe detector. The method includes: acquiring the raw signal collected by the quasi-hemispherical CdZnTe detector; normalizing the amplitude of the raw signal from the baseline to the highest point during the rise phase; dynamically selecting multiple energy thresholds; performing rise time discrimination on the normalized signal within each energy threshold; filtering out signals with rise times greater than RTflag and determining the rise time range corresponding to that energy threshold; constructing an intermediate energy spectrum corresponding to each energy threshold based on the signals with rise times greater than RTflag and the rise time range; and post-processing the intermediate energy spectrum to obtain the target energy spectrum corresponding to each energy threshold. This invention combines energy threshold discrimination, rise time discrimination, and energy spectrum post-processing algorithms to stably improve the energy resolution of the quasi-hemispherical CdZnTe detector for 662 keV gamma rays from the conventional 2%~3% to less than 1%.

The invention belongs to the technical field of drug detection, and particularly relates to a method for detecting lycium barbarum glycopeptide metabolites in rat plasma and tissues based on non-targeted metabonomics. According to the method, after rat lycium barbarum glycopeptides are administered through intragastric administration, blood, liver, intestine, lung and brain tissue samples after administration are collected. A UPLC-Q-TOF-MS (Ultra Performance Liquid Chromatography-Quadrupole Time of Flight Mass Spectrometry) analysis method is combined with a statistical analysis method comprising principal component analysis (PCA) and orthogonal-partial least squares discriminant analysis (OPLS-DA) to carry out in-vivo metabonomics analysis on the wolfberry glycopeptide. Then, based on a human metabolome database (HMDB), lycium barbarum glycopeptide differential metabolite is searched, so that the pharmacological action mechanism of the lycium barbarum glycopeptide is clarified from the perspective of metabolism, and a foundation is laid for deep research and development of the medicinal value of the lycium barbarum glycopeptide in future.

The invention belongs to the technical field of nondestructive testing of concrete structures, and relates to an urban underground comprehensive pipe gallery concrete engineering quality evaluation method which comprises the following steps: acquiring an original acoustic emission signal and performing structured processing to generate a structured acoustic emission event data object containing multi-dimensional feature information; carrying out feature correction and verification on the data object in combination with the environment data, and generating a focusing scanning instruction containing a target three-dimensional coordinate and a risk confidence coefficient; generating a parameterized adaptive focusing scanning configuration packet through physical parameter reverse mapping based on energy spectrum characteristics, driving a phased array ultrasonic system to perform targeted imaging on a target area and extracting acoustic image characteristics of defects; and finally, comparing the correlation between the data object and the image features, executing closed-loop verification, and adaptively updating discrimination parameters for risk determination. The technical problems of insufficient evaluation accuracy and verification lag caused by mutual separation of passive monitoring and active detection in the prior art are solved.

This utility model relates to the technical field of whisker production equipment, and provides a whisker collection device with purification function, including: a support and casters; a collection component, fixedly installed on the upper end of the support; a collection box, disposed at the bottom of the collection component; a feeding component, fixedly connected to one side of the collection component, and having a guide groove running through it; a dust removal component, fixedly disposed on both sides of the upper end of the collection component; a purification component, fixedly installed on one side of the feeding component; and a drying component, running through the back of the support. First, the device is moved to the unloading area, and all components are started. Whiskers enter the collection box through the hopper, dust is removed by a dust extraction fan, the sieve plate is screened by an anti-blocking mechanism, and the guide plate guides the whiskers to the guide plate before they fall into the material box. Simultaneously, the drying component dries the material, and some whiskers enter the purification chamber for purification. The purification liquid is recycled. This device integrates purification, screening, dust removal, and drying, comprehensively improving the quality of whiskers.

The invention belongs to the technical field of tunable lasers, and particularly relates to a grating scribed line density testing method and system based on a tunable laser, and the system comprises the tunable laser, a mechanical arm, a testing module and a control module. The characteristic that the tunable output range of a tunable laser (carrier) is sensitive to the grating scribed line density is fully utilized, and the mapping relation between waveguide coupling output multi-wavelength and the grating diffraction angle (or the displacement stride of an actuator) is deeply analyzed; according to the method, automatic analysis of the to-be-measured grating scribed line density under different batches, different substrate materials and different manufacturing process conditions is realized, and on the basis, accurate and rapid screening of qualified gratings suitable for tunable laser assembly is realized.

The utility model relates to a valve core end face length screening detection tool, and belongs to the technical field of detection tools. The problems of low detection efficiency and low measurement precision when the length of the end surface of the valve core is measured by manually holding a length measuring tool in the prior art are solved. The valve element end face length screening gauge comprises a mounting plate, a first detection plate and a second detection plate, the first detection plate and the second detection plate are oppositely mounted on the mounting plate in parallel, and a valve element detection cavity is formed between the first detection plate and the second detection plate; and one of the first detection plate and the second detection plate is mounted at different positions of the mounting plate relative to the other detection plate to form detection stations with different detection lengths. The screening gauge does not need to measure the length, so that the workload of manual measurement is reduced, and the detection efficiency is high. According to the invention, whether the to-be-detected valve core is a qualified product is determined by using the standard valve core end face length, so that a length measuring tool does not need to be held by a hand for measurement during detection, and low detection precision caused by reading of measurement personnel is avoided.

This invention provides a method for screening the stability of biomass shell-derived hard carbon raw materials. The method investigates the optimal ratio and relationship between the lignin content and sugar content measured after alkali treatment in biomass shell materials, thereby screening the stability of biomass shell materials from different sources and batches. When the lignin content of the biomass shell material itself and the sugar content measured after alkali treatment satisfy the relationship 15≤β×L≤60, an effective balance between the lignin and sugar content in the biomass shell material can be achieved, resulting in a good porous structure. This method provides a simple and rapid screening process to obtain biomass shell materials with better stability. When biomass shell materials with better stability are subsequently used to prepare hard carbon battery materials, the resulting hard carbon battery materials exhibit good overall performance in terms of cycle life, capacity, and coulombic efficiency.

This disclosure proposes a method for synthesizing multimodal data for decision-making attribution explanation in the mining industry, comprising the following steps: First, extracting causal logic from industry texts to construct a structured rule base; second, collecting on-site multimodal data and filtering out data pairs strongly correlated with rules through scenario-based weight modeling and cross-modal association mechanisms; third, using a multimodal generative adversarial network with rule constraint mechanisms and combined with rule mutation technology to synthesize multimodal data with complete logical chains and covering various complex working conditions; finally, ensuring data quality and achieving self-evolution of the knowledge base through a three-level verification and rule feedback mechanism, and storing qualified data in a hybrid storage database. This invention solves the problems of complex causal relationships and scarce high-quality training samples in the mining field, and can automatically and in batches generate logically rigorous and scenario-diverse multimodal data, improving the training efficiency and generalization ability of attribution decision models.

The application discloses a KASP molecular marker combination related to soybean salt tolerance and application thereof. Through soybean natural population salt tolerance phenotype identification and genome-wide association analysis (GWAS), three SNP sites extremely significantly associated with soybean salt tolerance traits are obtained, four salt tolerance candidate genes for regulating stress response and signal transduction of soybean are mined out; further based on the genomic sequence characteristics of the three SNP sites and the four candidate genes, seven molecular markers with high stability, good polymorphism and closely linked to salt tolerance traits are designed and verified to be developed. The application simultaneously provides a molecular marker development process, a soybean salt tolerance molecular identification method and application in germplasm screening and molecular assisted breeding. The application provides complete KASP primer combinations and molecular detection methods, which can provide efficient tools for accurate identification and screening of salt-tolerant soybean resources and molecular marker assisted breeding, and has a wide application prospect.

The invention discloses a construction method and application of a chlamydomonas reinhardtii source high-sweetness induced protein variant engineering algal strain. The preparation method comprises the following steps: firstly, modifying and optimizing protein coding sequences of sweet taste induction proteins Miraculin and Neoculin, so as to obtain optimized gene sequences of CrMiraculin and CrNeoculin; the method comprises the following steps: firstly, preparing a Chlamydomonas reinhardtii strain, then cloning the Chlamydomonas reinhardtii strain into an expression vector pGM6, constructing a recombinant expression vector, introducing the recombinant expression vector into the Chlamydomonas reinhardtii wild strain, and screening by using a paromomycin resistance marker carried by the vector to obtain a transformed strain. The obtained engineering algal strain can stably inherit and efficiently express the target sweet induced protein, and the chlamydomonas reinhardtii algal powder rich in the CrMiraculin and CrNeoculin sweet induced protein is obtained by performing fermentation culture on the engineering algal strain. Finally, through protein purification and mixing with acidic substances, sour taste can be induced to be converted into sweet taste, so that the method has great application potential in the fields of foods, seasonings or health-care products.

The application provides a fan clearance distance monitoring method and system based on multi-beam radar correlation detection. The system comprises a three-beam laser radar acquisition module, a data preprocessing module, a consistency verification module, an environment recognition and data compensation module, and a clearance value calculation and output module. The method comprises the following steps: S1, extracting echo intensity and slant range data of each beam; S2, performing time sequence filtering processing on the collected beam data and eliminating instantaneous jump points deviating from the mean value center; S3, performing multi-beam time correlation verification; S4, performing multi-beam numerical correlation verification; S5, calculating the visibility influence factor to determine the visibility environment; S6, fusing the effective data of the verification, using the remaining effective beam data for completion, and outputting the final fan clearance distance monitoring value. The application is applied to the technical field of wind power generation safety monitoring.

The invention belongs to the technical field of metal material quality test and evaluation, and discloses a primary magnesium ingot internal inclusion level evaluation method and system.The method comprises the steps that firstly, the hydrogen evolution rate and the characteristic low-frequency impedance modulus value of a primary magnesium ingot to be tested are obtained; and then the level of the inclusions in the primary magnesium ingot to be tested is evaluated through the hydrogen evolution rate, the characteristic low-frequency impedance modulus value and the pre-constructed hydrogen evolution-impedance collaborative response criterion, and the method is easy and convenient to operate, short in test period, low in cost and high in accuracy. The method can be used for rapidly identifying the extremely inferior magnesium ingot with components meeting requirements and a large number of densely distributed inclusions inside without depending on microscopic analysis, is suitable for quality monitoring and supplied material acceptance in the production process, and has remarkable industrial practical value.

The invention discloses a casting powder slag entrapment resistance evaluation method, device and equipment, a medium and a product. The method comprises the following steps: acquiring a spreading image of to-be-detected casting powder after complete melting on a sample carrier, then determining a spreading area of the to-be-detected casting powder based on the spreading image, finally acquiring a reference spreading area, and comparing the reference spreading area with the spreading area to obtain the slag entrapment resistance of the to-be-detected casting powder, the reference spreading area is obtained by measuring the reference casting powder sample with the viscosity being the preset viscosity value and the slag entrapment defect occurrence rate being lower than the preset threshold value under the same test condition as the to-be-detected casting powder, and the slag entrapment resistance of the casting powder can be judged in advance before production, so that resource waste is reduced, and the productivity is improved.

The invention belongs to the field of separation and filtration, and particularly relates to a filtering membrane with aperture sizes in gradient arrangement and a preparation method thereof. The preparation method comprises the following steps: (1) preparing a membrane casting solution: mixing and dissolving dimethylacetamide, polyether sulfone, polyethylene glycol and an auxiliary agent, and defoaming to obtain the membrane casting solution; (2) preparing core liquid: mixing dimethylacetamide with water to obtain the core liquid; (3) preparing a gel bath: mixing dimethylacetamide with water to obtain the gel bath; the concentration of the gel bath is smaller than that of the core liquid; and (4) preparing the filter membrane: enabling the membrane casting solution to be in contact with the core solution and the gel bath through a spinneret plate, and enabling membrane filaments to pass through the gel bath to a filament collecting tank, so as to obtain the filter membrane of which the pore sizes are arranged in a gradient manner. The pore diameter of the filtering membrane is changed from a large pore diameter on one side to a small pore diameter on the other side in a gradient manner, and the filtering membrane has high retention rate and high flux, high mechanical strength, easy cleaning and low energy consumption.

The utility model relates to the technical field of tea stalk crushing, and discloses a tea stalk carbonized material crushing machine which comprises a feeding port, a protective shell is fixedly connected to the bottom of the feeding port, a first motor is fixedly connected to the rear side of the protective shell, and a rotating rod is rotationally connected to the front side of the first motor. A plurality of conveying blades are fixedly connected to the outer wall of the rotating rod, a shell is fixedly connected to the bottom of the protective shell, a second motor is fixedly connected to the rear side of the shell, a rotating shaft is rotatably connected to the front side of the second motor, a plurality of blades are fixedly connected to the outer wall of the rotating shaft, and a discharging port is fixedly connected to the bottom of the outer wall of the shell. The feeding port and the protective shell ensure that materials enter the machine safely, the first motor drives the rotating rod and the material conveying blades to efficiently convey the materials to prevent blockage, the second motor drives the rotating shaft and the blades to rapidly rotate to smash the materials, the filter screen screens the smashed materials, and the smashing effect and efficiency are improved.

The invention provides a customer screening apparatus, a customer screening method, and a computer program product. The customer screening device for performing customer screening based on the game adversarial network comprises a recessive code generation module which generates a recessive code vector based on pre-prepared data of a target customer, and the recessive code vector maps features which are meaningful to judgment of the target customer; the virtual customer information generation module is used for generating virtual customer information based on the implicit coding vector and the random noise vector by utilizing a generative network in the game adversarial network; the game confrontation module is used for discriminating the virtual customer data based on the real customer data by using a confrontation network in the game confrontation network; and the auxiliary network module is used for providing training feedback for the generative network by using an auxiliary neural network according to the implicit coding vector and the virtual customer data so as to improve the correspondence between the generated virtual customer information and the implicit coding vector, and alternately training the generative network and the adversarial network based on a loss function so as to improve the correspondence between the generated virtual customer information and the implicit coding vector. The loss function comprises mutual information operation between a nonlinear difference function of the generative network and the adversarial network and an output function of the generative network and an output function of the auxiliary neural network.

The embodiment of the application relates to the battery technical field, and discloses a battery pack equalization method, device and electronic equipment, the method comprises the following steps: obtaining cloud equalization identification information sent by a network device; wherein the cloud equalization identification information is generated by the network device based on historical charging data and / or historical discharging data of a target battery pack, and is identification information indicating whether each single battery in the target battery pack needs equalization; generating local equalization identification information based on current state parameters of the target battery pack; the local equalization identification information is used for indicating whether each single battery in the target battery pack needs equalization; based on the cloud equalization identification information and the local equalization identification information, determining a target single battery needing equalization in the target battery pack, and equalizing the target single battery.

This invention provides a water surface photocatalytic oxidant release device, belonging to the field of photocatalytic equipment technology. It includes a main structure with an internal reaction chamber. The chamber has an inlet and an outlet at both ends, respectively, both communicating with the reaction chamber. An oxidant release membrane, a sulfonated polyethersulfone selective permeation membrane with a pore size <1 nm, is located at the outlet. The reaction chamber contains a catalyst support and a light source system. The catalyst support is made of porous ceramic and loaded with a g-C3N4 catalyst. The light source system includes an LED board fixed in the center of the catalyst support and surrounded by it. The LED board has a ring-shaped array of blue LED beads emitting light of a specific wavelength of 450-470 nm. An energy and control compartment is located inside one side wall of the main structure, housing a battery pack and a control unit board. This invention can significantly improve environmental benefits and treatment efficiency, and greatly enhance long-term operational stability and economic efficiency.

This invention relates to the field of sodium-ion battery technology, specifically to a rapid screening method for low-temperature sodium-ion batteries and its application. The method involves assembling positive electrode, negative electrode, separator, and electrolyte into different sodium-ion batteries, followed by charging and formation. Q-V data for each sodium-ion battery during the formation process are obtained, and dQ / dV-V curves are plotted to determine the peak position and area of ​​the redox peak. After formation, the sodium-ion batteries undergo aging and capacity testing to obtain finished sodium-ion batteries. Under low-temperature conditions, the finished sodium-ion batteries are subjected to DCR testing to obtain their low-temperature DCR resistance. Based on the peak position and area of ​​the redox peak and the low-temperature DCR resistance, the low-temperature performance of different sodium-ion batteries is determined. This method can quickly and accurately screen sodium-ion battery systems with relatively superior low-temperature performance, shortening the cell low-temperature performance testing cycle and improving the screening efficiency of low-temperature sodium-ion batteries.

This invention belongs to the field of cross-industry information model construction technology, and relates to a method and system for semantic construction and automatic generation of multi-industry information models. The method includes: generating standard format requirement parameters for cross-industry models based on cross-industry model construction requests; decomposing the standard format requirement parameters into industry requirement parameter sets; generating customized meta-models by combining the industry requirement parameter sets, a meta-model base library, and a cross-industry feature library; semantically aligning the customized meta-models based on an industry terminology list and a cross-industry semantic rule library to obtain a cross-industry model element semantic alignment table; constructing lightweight model structures based on the cross-industry model element semantic alignment table; performing syntax validation on each lightweight model structure; and generating standardized model files corresponding to each target industry. This invention achieves automated construction of cross-industry models from scratch, improves model adaptation success rate and semantic accuracy, and avoids the problem of cross-industry models being unable to communicate with each other.

The invention relates to the technical field of alloy materials, and discloses an alloy material component screening method, system and equipment and a medium. The method comprises the following steps: constructing an alloy component performance data set; constructing a multi-model performance prediction model, and training the multi-model performance prediction model according to the alloy component performance data set to screen out a performance prediction reference model; performing global search according to a genetic algorithm and the performance prediction reference model to obtain an alloy component scheme and a corresponding performance prediction result; according to the alloy component scheme and the performance prediction result, weighted assignment sorting is carried out to screen out a target alloy combination; and performing simulation verification and process analysis on the target alloy combination to optimize the performance prediction reference model. According to the method, deviation of a single model is avoided through a multi-model performance prediction model, the genetic algorithm is introduced to ensure global search and avoid local optimization, the closed-loop process design is formed by combining simulation verification and process analysis, the target alloy combination meeting the requirements is rapidly and accurately screened, and the development efficiency is improved.

The application discloses a power transformer winding fault detection method based on vibration spectrum analysis and relates to the technical field of power equipment detection, which comprises the following steps: collecting original vibration acceleration signals of a transformer and preprocessing the original vibration acceleration signals to form a signal frame sequence; applying an improved Hilbert-Huang transform to each signal frame to obtain a set of intrinsic mode functions, calculating a component instantaneous frequency sequence and extracting a frequency band with an energy proportion exceeding a preset threshold as a candidate fault characteristic frequency band; extracting a subband signal of each frequency band and calculating a high-order statistical quantity feature matrix; loading a health state reference feature matrix matched with a transformer model, capacity and voltage grade; comparing the two in an element-by-element manner to identify an abnormal feature submatrix; spatially clustering the abnormal feature submatrix to form a suspected fault mode cluster; and outputting a specific winding fault mode diagnosis result. The method can accurately extract fault characteristics and improve the accuracy and reliability of fault detection.

The invention relates to the technical field of thermoelectric Internet of Things safety, in particular to a heat meter verification data intelligent analysis system based on heat supply multi-parameter simulation, which comprises a multi-parameter working condition simulation module, a data acquisition and communication module, a verified household heat meter, a data processing and storage module, an intelligent analysis core module and a calibration optimization suggestion module, the multi-parameter working condition simulation module accurately reproduces water quality, temperature, flow velocity and battery electric quantity coupling working conditions; the data acquisition and communication module realizes unified acquisition of heat meter data of multiple brands; the intelligent analysis core module quantifies working condition coupling influence, sorts heat table performance and predicts the service life of the battery through a multivariate regression + SHAP hybrid model, an entropy weight-TOPSIS method and an ARIMA model; the calibration optimization suggestion module generates a dynamic calibration coefficient. According to the invention, all-condition verification, intelligent analysis and accurate optimization of the heat meter are realized, data support is provided for purchase, calibration, operation and maintenance of the heat meter of a heat supply enterprise, and the metering accuracy and the energy utilization efficiency are improved.

This invention relates to the field of data governance and big data processing technology, and discloses a data lineage confidence verification method, comprising: asynchronously collecting metadata from a multi-source heterogeneous system, hierarchically parsing and constructing an initial lineage graph; calculating the comprehensive confidence score of each lineage edge based on multi-dimensional feature indicators and a confidence assessment model, and filtering low-confidence edges to form a list to be verified; performing at least two collaborative verification methods among cross-source active detection, AI semantic reasoning, and manual visual annotation on the lineage edges to be verified, and triggering them sequentially; writing the verification results back to the graph, updating the topology and confidence state, and iteratively executing scoring, filtering, and verification until a preset closed-loop termination condition is met. This method can achieve high-precision, interpretable, and self-optimizing lineage governance.

The present application relates to the technical field of large model, especially to a reading recommendation large model generation method and system based on big data resources.The method comprises the following steps: obtaining a text to be evaluated and collecting reading materials of multiple learning stages and multiple types, labeling them, aggregating and verifying the consistency of the labeling results, and constructing a verified recommendation database; based on the recommendation database, performing low-rank adaptive fine-tuning on the general pre-training large model to obtain a domain model, and monitoring and correcting the performance and bias during the training process; performing multi-dimensional analysis on the text to be evaluated to generate a text feature vector, combining semantic vector similarity retrieval to perform hybrid retrieval on the recommendation database; the present application realizes more accurate and efficient reading recommendation through the construction of the reading recommendation large model.

The utility model relates to the technical field of rice noodle screening, and provides a screening device for rice noodle production, which comprises a shell, a main rod is rotatably connected in the shell and extends out of one end, two driven cranks are fixedly sleeved on the outer surface of the main rod, and one end of each driven crank far away from the main rod is rotatably connected with a connecting rod. And shaft blocks are rotationally connected to the ends, away from the driven cranks, of the connecting rods, adjusting plates are fixedly connected to the tops of the two shaft blocks, wedge blocks are fixedly connected to the two sides of each adjusting plate, and wedge grooves are formed in the two sides of the inner surface of the shell. The rice flour screening device is provided with a vibration amplitude adjustable structure, meets screening requirements of different production stages and different material characteristics, achieves more accurate screening by flexibly adjusting vibration intensity, is suitable for screening rice flour which is hard in texture and not prone to breakage, enables the rice flour to rapidly pass through a screen mesh, reduces fine flour residues, and improves screening efficiency. Soft or fine rice flour is processed, and the yield is increased.