170 results about "Sample pool" patented technology

A method for detecting circularity of silicon carbide micro powder relates to the technical field of silicon carbide detection, and comprises the following steps: S1, turning on a flowing particle imager, preheating and then carrying out background detection; s2, adding standard particles into the sample pool, and enabling the flowing particle imager to perform automatic focusing; s3, adding the prepared silicon carbide micro powder sample into a sample pool, and detecting by using a flowing particle imager; and S4, storing the detection result, checking the silicon carbide micro powder particle image, and recording the detection result. The preparation of the silicon carbide micro-powder sample comprises the following steps: S31, adding the dried silicon carbide sample into a dry and clean beaker; s32, sequentially adding an impregnating compound, a dispersing agent and deionized water into the beaker; s33, carrying out ultrasonic dispersion; and S34, taking out the uniformly mixed sample. The method can solve the technical problem of detection failure caused by easy agglomeration of the silicon carbide sample when the flowing particle imager is used for detecting the circularity of the silicon carbide micro powder.

The invention relates to the technical field of fluorescence detection, in particular to a three-dimensional fluorescence detection correction method based on double light sources, and the correction method comprises the following steps: obtaining double light source optical parameters, sample cell geometric parameters and original fluorescence intensity data, the geometric parameters of the sample pool comprise the length, width, height, pool wall reflection coefficient and sample attenuation coefficient of the sample pool; and constructing a light intensity distribution model in a three-dimensional space of the sample cell based on the optical parameters of the double light sources and the geometric parameters of the sample cell. According to the invention, accurate characterization of the light intensity of any spatial point in the sample cell is realized, light intensity deviation caused by various factors is quantified from the source, personalized correction is carried out on each spatial point, the influence caused by optical path difference at different positions, light source fluctuation and element error is effectively offset, and the accuracy and accuracy of the detection result are improved. The consistency of detection values of standard substances with the same concentration in different areas of the sample pool is greatly improved, and the linearity of a standard curve is remarkably optimized.

The invention discloses a security detection method and device for network transmission data, and belongs to the field of data security. The method comprises the following steps: performing security detection on network transmission data by using a security detection model and a clustering model; and outputting a detection result based on comparison of detection results of the security detection model and the clustering model. When the comparison result is inconsistent, calculating a comprehensive safety coefficient, and storing the network transmission data into an experience sample pool; and respectively carrying out model optimization on the security detection model and the clustering model by utilizing samples in the empirical sample pool. According to the scheme, the supervised security detection model and the unsupervised clustering model are used for parallel detection, and bimodal high-precision detection is generated; through cross-modal conjoint analysis of two modal models, doubtful network transmission data is stored in an experience sample pool, parallel optimization of a supervised security detection model and an unsupervised clustering model is realized by using a sample of the experience sample pool, online closed-loop hot update of the models is realized, dynamic perception of novel attacks is improved, and the misjudgment rate is reduced.

This invention discloses a refractive index detection sensing device, comprising a support platform. A sample cell is adjustablely mounted at the center of the top of the support platform, and the sample cell contains a 3C-SiC nanocore-shell structure and a background medium. An incident unit is adjustablely mounted on one side of the support platform relative to the sample cell, and a detection unit is adjustablely mounted on the other side of the support platform relative to the sample cell. The 3C-SiC nanocore-shell structure comprises a 3C-SiC nanosphere core and a shell medium enclosing the 3C-SiC nanosphere core. The radius of the 3C-SiC nanosphere core is a, and the radius of the shell medium is b. The 3C-SiC nanocore-shell structure is embedded in a background medium with a uniform relative permittivity. This invention exhibits a detection sensitivity under a nonlocal model that is more than ten orders of magnitude higher than that under a local model, and can accurately capture minute changes in refractive index on the order of 0.01.

The invention belongs to the field of radar automatic target recognition, and relates to an SAR target recognition method and device based on projection features and improved self-learning. The method comprises the following steps: respectively processing each sample parameter of a training sample by a classifier T respectively trained by each prototype set to obtain integrated projection features of the sample parameters; initializing an SVM classifier based on the labeled sample feature set; taking out part of samples from the unlabeled sample feature set, and constructing a temporary sample pool; classifying the samples in the temporary sample pool by using an SVM classifier; evaluating the classification confidence of the samples in the temporary sample pool by using the classification posterior probability; adding a plurality of samples with high confidence into the labeled sample feature set, and retraining the SVM classifier; repeating for multiple times until an optimal SVM classifier is obtained; and target identification is carried out based on the optimal SVM classifier. The target recognition performance under the condition of a small number of labeled samples and a large number of unlabeled samples is effectively improved.

The application discloses a kind of power material evaluation sample dynamic construction method and device, method includes: obtaining historical original evaluation file and carrying out format conversion, obtain structured initial evaluation sample pool;The entropy of each initial evaluation sample in initial evaluation sample pool is analyzed and is filtered, and evaluation sample set is obtained;Initial evaluation sample in evaluation sample set is standardly marked, and is combined with pre-constructed thought chain model to carry out thought chain analysis, and obtain standard evaluation triplets;According to preset evaluation reward model, standard evaluation triplets are evaluated and screened, and power material evaluation sample library is generated;In response to real-time original evaluation file, the update of power material evaluation sample library is completed, and the dynamic construction of power material evaluation sample library is completed.Based on the construction of power material evaluation sample library, it is updated to ensure the quality of evaluation sample in power material evaluation sample library, and provide good data basis for model training.

The application discloses a system and method for detecting SF6 gas decomposition products based on a spectrum method, and the system comprises an SF6 gas and decomposition product collecting device, a gas sample pool, a high-purity argon gas bottle, a laser device, an argon gas collecting pool, a spectrum analysis and detection device and a gas recharging device; the SF6 gas and decomposition product collecting device, the high-purity argon gas bottle, the laser device, the argon gas collecting pool, the spectrum analysis and detection device and the gas recharging device are all connected with the gas sample pool; the gas sample pool is embedded with a one-side wideband total reflection mirror and a quartz window sheet, the wideband total reflection mirror is symmetrically provided with a first arc-shaped total reflection mirror and a second arc-shaped total reflection mirror which are perpendicular to the wideband total reflection mirror, and the quartz window sheet is opposite to the wideband total reflection mirror; and the spectrum analysis and detection device collects scattered light through the quartz window sheet. The application has high detection sensitivity, fast detection speed and can realize real-time online analysis and detection.

This application discloses a method, apparatus, and device for optimizing a large model targeting a specific vertical domain, belonging to the field of artificial intelligence. The method includes: acquiring a sample pool, which includes general domain data and vertical domain data; constructing current-round training data based on the sample pool, and training an initial large model based on the current-round training data to obtain a target large model; constructing test questions for the vertical domain data based on the technical standards of the target vertical domain; generating response results for the test questions using the target large model; determining domain knowledge blind spots from the vertical domain data based on the understanding of the technical standards based on the response results; updating the current-round training data based on the domain knowledge blind spots to obtain the next-round training data, and training the target large model again based on the updated next-round training data.

The invention provides a large model entity alignment prompting method and system. The method comprises the following steps of: obtaining semantic embedding representation by utilizing joint embedding and fine tuning of a knowledge graph and a text corpus; retrieving a text block set related to the semantics of the query entity in a text corpus through a k-nearest neighbor algorithm retrieval technology to obtain a retrieval result; searching k neighbor entity pairs of the query entity pair from an expert labeled sample pool based on the key information, and taking the k neighbor entity pairs as positive samples; searching k neighbor entity pairs of the query entity pair from a public data set based on key information as supplementary examples; constructing the positive samples and the supplementary samples into a demonstration sample set; converting the demonstration sample set into a natural language text to obtain a sample text representation; and inputting the retrieval result and the sample text representation into the large model as an entity alignment prompt. According to the large model entity alignment prompting method and system, the reasoning ability of the large model in the entity alignment task is remarkably improved.

This disclosure relates to the field of reliability technology, specifically to a method for global sensitivity analysis of a radome structure. The method includes: obtaining important sampled samples of the input variables of the radome structure and a sample pool of the distribution parameters of the input variables; obtaining a training set for a Kriging failure model of the radome structure based on the important sampled samples, and training the Kriging failure model using the training set; when it is determined that each sample in the sample pool meets preset conditions, substituting the samples of the distribution parameters into the Kriging failure model to obtain the global sensitivity analysis result of the radome structure. This disclosure can improve the performance and reliability of radome structures.

A new intelligent drilling sample pool of downhole equipment for drilling sample pool, including sample pool main body, upper impeller system, lower impeller system and detection device, sample pool main body upper end installs central processing unit, central processing unit control signal control element in device, the liquid passage is longitudinally opened in the sample pool main body, the liquid passage upper side extends upward, the liquid passage upper side is sample pool entrance, the liquid passage lower side is sample pool outlet, the upper layer horizontal left side of sample pool main body is opened upper impeller cavity, upper impeller system is horizontally installed in upper impeller cavity;The upper layer horizontal right side is opened piston cavity, piston cavity is communicated with side wall, detection device is installed in piston cavity upper end, lower impeller system is installed in lower impeller cavity, the intelligent detection of central processing control device signal control element is carried out in the application, the problem of difficult extraction of uplink oil-based drilling fluid is solved, the distance difference and time difference problem of sample liquid from downhole to ground and then detection is optimized, and drilling efficiency is improved.

The invention discloses a portable Raman liquefied natural gas composition analysis device and method. The portable Raman liquefied natural gas composition analysis device comprises a sample pool, and the sample pool comprises a sample pool inner container, a sample pool vacuum outer container and a heat preservation layer which are sequentially arranged from inside to outside in a sleeving mode; a sample circulating cooling inner cavity is formed in the sample cell liner; a gap is formed between the sample cell inner container and the sample cell vacuum outer container to form a vacuumizing heat preservation outer cavity; a Raman probe insertion pipeline protruding outwards is arranged above the sample pool, and an insertion opening is formed in the top of the Raman probe insertion pipeline; a gas inlet pipeline is arranged in the sample pool, and one end of the gas inlet pipeline extends out of the sample pool and is used for being connected with an LNG pipeline. By designing the special LNG sample pool, LNG liquid sampling can be carried out, and the problem that an LNG station cannot directly measure the composition of the LNG liquid by adopting a Raman spectrum due to lack of a Raman interface is solved.

The invention discloses a natural circulation system reliability robust optimization design method based on active learning. The method comprises the following steps: determining a system limit state function and screening key parameters; establishing a mean value sensitivity model and constructing a probability optimization model; adopting multi-objective optimization to obtain a Pareto solution set; quantizing parameter uncertainty and constructing an initial proxy model; performing reliability evaluation based on the initial agent model and generating a candidate sample pool; iteratively updating the proxy model through an active learning function; reliability analysis is carried out again based on the updated model, and the reliability of the solution set is judged; and performing optimization decision on the solution set meeting the requirements to output an optimal solution. According to the method, the agent model is adaptively updated through the active learning strategy, the evaluation efficiency is remarkably improved while the calculation precision is ensured, collaborative optimization of the system weight and the robustness under the reliability constraint is realized, and the anti-interference capability of the natural circulation system facing uncertain factors is effectively improved.

The invention discloses a portable microfluid bubble manual removal device which comprises a first exhaust part, a second exhaust part, a sample pool and an ultrasonic bath, the sample pool is filled with a solution, and the sample pool is arranged in the ultrasonic bath; the first exhaust part comprises a first injector, a first pipeline and a first air collecting bag; the second exhaust part comprises a second injector, a second pipeline and a second air collecting bag; the first air collecting bag and the second air collecting bag each comprise a hollow bag cavity and a shell, each hollow bag cavity comprises an upper bag cavity part, a middle bag cavity part and a lower bag cavity part which are sequentially connected, and the upper bag cavity parts, the middle bag cavity parts and the lower bag cavity parts are of an integrated structure. In the structural design of the first air collecting bag and the second air collecting bag, the diameter of the middle bag cavity part obtained when the diameter of the middle bag cavity part serves as the characteristic length and the diameter of the middle bag cavity part obtained when the liquid level height in the hollow bag cavity serves as the characteristic length are met at the same time, and it is ensured that bubbles in a sealed fluid loop can be effectively discharged; and bubbles on the wall of each component of the fluid loop are eliminated.

The invention provides a knowledge graph entity alignment method and system based on large model labeling and active learning. The method comprises the following steps: an expert searches corresponding entities according to entity characteristics of entities in a representative entity set to obtain positive example entity pairs; based on the positive example entity pair, utilizing a negative sampling strategy and a dynamic deletion strategy to obtain a negative example entity pair and a difference entity pair, and constructing the positive example entity pair, the negative example entity pair and the difference entity pair into a sample pool; labeling the entities and the entity pairs in the sample pool by using a large model to obtain a labeling result; evaluating the labeling quality of the labeling result through an evaluation index; and generating a candidate set by using a KNN algorithm based on the sample pool and the labeling quality. According to the knowledge graph entity alignment method and system based on large model labeling and active learning, the problem that an entity alignment model depends on manual labeling is solved, and the accuracy and reliability of entity alignment are remarkably improved.

The invention discloses a chest radiograph report generation method based on thinking chain reasoning and active learning. The method comprises the following steps: collecting a multi-view chest radiograph image and a corresponding report; performing view angle identification and view angle classification on the multi-view-angle chest radiograph image; a shared visual encoder is adopted to extract sequence features of all visual angles, and a multi-visual-angle visual token is generated through cross-visual-angle feature fusion; constructing a systematic hierarchical diagnosis process, sequentially executing structured five-stage chain reasoning for each anatomical structure in the multi-view chest radiography image, and generating a structured reasoning prompt; inputting the multi-view visual token and the structured reasoning prompt into a large language model, and generating a target sequence through an autoregression mode; constructing a scoring function based on joint scoring of uncertainty and inconsistency; labeling the samples in the unlabeled sample pool according to the sample scores in each round of iteration; and inputting a to-be-tested multi-view chest radiograph image into the trained language model, and generating a thinking chain and a complete report text.

The invention provides a laser absorption and LIF imaging fused water quality COD detection method and system, and belongs to the technical field of water quality monitoring, the method comprises the following steps: placing a to-be-detected water sample in a sample pool, and vertically irradiating the water sample with laser; performing high-concentration COD detection by adopting an ultraviolet absorption spectrophotometry based on a Lambert-Beer law in combination with a calibration curve; a laser induced fluorescence (LIF) imaging technology is adopted, the RGB gray value of the ROI of a fluorescence image is extracted, and low-concentration COD detection is carried out in combination with a fluorescence intensity-COD model; automatically matching and detecting according to the concentration of the water sample, and outputting a final COD detection value; the system comprises a laser emission module, an image acquisition module, a sample detection module, a transmission light detection module and a data processing module. The advantages of two optical detection technologies are fused, rapid and accurate detection of COD of various water bodies is achieved, the system sensitivity is high, the structure is compact, and real-time monitoring of COD of the water bodies is achieved.

This invention discloses an adaptive PC-Kriging reliability analysis method and system based on active learning. The method includes: using weighted clustering to obtain uniformly distributed first candidate samples from a pre-generated MC sample pool and constructing an initial PC-Kriging model; using interval reduction to select samples and construct a new sample pool, dividing the new sample pool into two sub-sample pools: a safe domain and a failure domain; for each sub-sample pool, weighted clustering is used again to select uniformly distributed second candidate samples; using the second candidate samples distributed in the failure domain and the safe domain, crossing points are constructed; the constructed crossing points are used as new experimental points to iteratively update the PC-Kriging model; convergence is determined and the results are output. The technical solution of this invention focuses on key regions through interval reduction and updates the model based on crossing points, which can more accurately approximate the limit state surface, reduce the number of experimental points, improve prediction accuracy, and is more efficient in utilizing computational resources.

This invention discloses an active learning method for semantic segmentation of remote sensing images of road cracks based on a self-supervised dual representation mechanism, belonging to the field of computer vision and intelligent interpretation technology of remote sensing images. The method includes: extracting global semantic distribution features of unlabeled sample pool images using a first self-supervised model with frozen parameters; calculating pixel-level reconstruction error maps using a second self-supervised model with frozen parameters; generating a comprehensive sampling value score by weighted fusion of the two types of self-supervised features, with a cold-start hybrid sampling strategy adopted in the first iteration and dynamic adjustment of fusion weights in subsequent iterations; selecting a high-value sample set from the unlabeled sample pool and performing pixel-level annotation; and independently training the semantic segmentation main network using the labeled sample set, with the self-supervised model not participating in gradient updates throughout the process. This invention achieves complete decoupling between the sampling module and the segmentation model, significantly improving crack edge segmentation accuracy and model convergence efficiency with low annotation costs.

This invention provides a gas detection method based on non-dispersive infrared absorption spectroscopy, comprising the following steps: (A1) Measurement light emitted from a light source passes through a chopper, the rotation speed of which changes periodically; after passing through the chopper, part of the measurement light passes through a sample cell and enters a first container filled with the gas to be measured, and part of the measurement light passes through a reference cell filled with a gas containing zero content of the gas to be measured and enters a second container filled with the gas to be measured; a conversion unit senses the periodic changes in gas pressure in the first and second containers and outputs electrical signals; zero gas and standard gas of the gas to be measured are respectively introduced into the sample cell, and the conversion unit outputs electrical signals V0(t) and V... span (t), where t is time; (A2) Fit the function F0(t) = V span (t)-V0(t); (A3) The sample gas to be tested is introduced into the sample cell, and the conversion unit outputs an electrical signal V(t); (A4) The fitting function F(t)=V(t)-V0(t)=A·F0(t)+B, where A and B are the fitting parameters; (A5) The concentration of the gas to be tested in the sample gas is obtained as C=A·C span C span This refers to the concentration of the gas being measured in the standard gas. This invention has advantages such as small detection error.

The application discloses a kind of fast simulation method of ion scattering track cluster driven by bypass factor, including constructing basic radiation parameter table;According to the basic radiation parameter table, search discrete energy point, form lookup table;According to the variation of bypass factor of each discrete energy point in lookup table, determine the scattering angle distribution characteristic parameter corresponding to discrete energy point, and generate scattering angle sample based on scattering angle distribution characteristic parameter and store in scattering angle sample pool;Initialize three-dimensional grid, set the initial position and direction of incident ion;Carry out incident ion transport and energy deposition cycle;Parallel execution of the simulation task of multiple incident ions, respectively in independent three-dimensional grid copy incident ion transport and energy deposition cycle, finally all three-dimensional grid copy is obtained Total dose distribution.This method achieves optimal balance between calculation efficiency and accuracy, so as to generate high-fidelity three-dimensional dose distribution chart.

The invention relates to the technical field of chips, in particular to a chip which is sequentially provided with a first centrifugal area, a second centrifugal area, a third centrifugal area and a detection area. Wherein a sample adding pool and a first pool are arranged in the first centrifugal area; the second centrifugal area is internally provided with a first quantitative pool and a second quantitative pool, wherein the first quantitative pool is connected with the sample adding pool through an I-type channel; the third quantitative tank is connected with the first tank through a channel I; a mixing cavity is arranged in the third centrifugal area, the sample and the buffer solution synchronously enter the mixing cavity under the action of centrifugal force to form a mixed solution, and the mixed solution reacts with a target reagent in the mixing cavity. The reaction efficiency can be improved, and the accuracy of a detection result is improved.

Provided are a cell imaging device and a cell imaging method that can shorten the time period of taking images of cells in a liquid sample, compared with conventional techniques. This cell imaging device 100 introduces a urine sample containing cells into an internal space 11 of a sample cell 10, moves in a second direction at least one of the sample cell 10 and an objective lens 52 while at least one of the sample cell 10 and the objective lens 52 is moved in a first direction, the second direction being different from the first direction, and takes, at a plurality of imaging positions, images of cells contained in the urine sample by means of an imaging unit 50.

The application discloses a kind of cultivation of smell and headspace analysis combination's odd nan hand string nondestructive identification method, it is related to precious timber true and false identification and quality detection technical field, including following step A: sample pool data model establishment: a1, sample selection: collect the cultivation of different producing areas, gram weight, shape and size odd nan hand string genuine sample not less than 200;Collect by high pressure injection paste, press oil, bubble medicine, high throwing, immersion paste smearing and non-lignum aquilariae imitated fake sample not less than 200;A2, electronic nose data acquisition: using electronic nose to the sample carries out smell data acquisition, acquisition condition includes;By being provided with electronic nose data acquisition and key sensor screening, the whole identification process does not need to be destructively treated to sample any cutting, grinding etc., completely maintain the integrity and value of hand string, and electronic nose detection can be completed within 1 hour, headspace mass spectrum analysis as supplementary means can be completed within 3 hours, improve detection efficiency.

The present disclosure provides an audio recall method, a model training method, a device and an electronic device, relates to the technical field of cloud computing, in particular to the technical field of deep learning, intelligent search and voice technology, and the audio recall method comprises: acquiring a first audio; segmenting the first audio to obtain N audio segments, any two adjacent audio segments in the N audio segments partially coincide, and N is an integer greater than 1; recalling a second audio corresponding to each of the N audio segments from a sample pool.

The utility model discloses a digital micro-fluidic system containing a quartz crystal microbalance sensor. The digital micro-fluidic system comprises an integrated system and a main body, the integrated system comprises a constant temperature system, a microfluidic system, a quartz crystal microbalance system, a control system and a data processing system, automatic control of subsystems of the instrument is realized by using an embedded system, and integrated control of the subsystems is completed by combining system programming. A main body is designed to surround a micro-fluid channel of a micro-fluidic chip and a layout of a quartz crystal sensor, the QCM sensor is connected with the micro-fluid channel of the micro-fluidic chip, it is ensured that fluid can be in contact with the sensor through the channel, fluid flowing is accurately controlled through a peristaltic pump and the micro-fluid channel, and a sample pool and a waste liquid pool are arranged; the container is used for storing samples and collecting waste liquid; mechanical support design is carried out on an instrument support structure, and internal structure protection design is carried out on each control and interface unit.

The invention discloses a category rebalance overflow state identification method based on underground imbalance monitoring data, and relates to the technical field of petroleum drilling engineering and intelligent well control. The method comprises the following steps: acquiring multi-source underground monitoring data generated in a drilling process, and constructing an initial labeled sample set and an unlabeled sample pool; a joint self-supervision feature learning model is constructed, and depth characterization of drilling features is realized by introducing a feature extraction network based on an attention mechanism and potential feature space modeling; under an active learning framework, performing category grouping on unlabeled samples according to a model prediction result, and adaptively allocating a sample query number through a category rebalance quota optimization strategy; and further introducing an uncertainty estimation and mutual information measurement method based on random reasoning, preferentially selecting samples with relatively high information amount for manual annotation, and carrying out iterative updating on the model. According to the method, the problems of extremely unbalanced overflow sample distribution and high manual labeling cost are effectively relieved, the recognition precision and the model generalization ability of minority overflow events are remarkably improved under the condition of limited labeling budget, and the method has good engineering application value.

The application relates to the technical field of natural language processing, in particular to an AI generated text detection method and system based on a deep learning model. The method comprises the following steps: performing character normalization and sentence segmentation on a to-be-detected text, segmenting the text by a sliding window to generate a segmented sequence; constructing deep semantic discrimination features, logarithmic probability difference trace features, style structure features and entity consistency features; inputting a gate fusion discrimination network, outputting a generated probability and confidence degree after probability calibration, and outputting a segmented contribution degree; performing threshold adaptive determination according to a risk level, a text length and the confidence degree, outputting a review state and generating a traceable record when the confidence degree is insufficient; and triggering incremental updating based on drift monitoring and a feedback sample pool. The application improves cross-domain robustness and interpretability, reduces false positives and false negatives, and supports long-term stable operation.

This invention relates to an active learning-based optical remote sensing target detection method, comprising: acquiring a sample set of optical remote sensing target images; constructing a test sample set and an unlabeled sample pool; randomly selecting a batch of samples from the unlabeled sample pool for expert annotation as an initial training sample set; building a single-stage optical remote sensing target detection model; designing an active learning sampling strategy; training the single-stage optical remote sensing target detection model using the training sample set to obtain an initial optical remote sensing target detection model; testing the optical remote sensing target detection model using the test sample set to determine whether the model has reached the training termination condition; if not, sampling from the unlabeled sample pool according to the active learning sampling strategy, adding the samples to the training sample set after expert annotation, and continuing iteratively training the optical remote sensing target detection model until the training termination condition is reached. This method can effectively reduce sample annotation costs while maintaining the model's detection performance.