45 results about "A priori probability" patented technology

The purpose is to appropriately manage abnormal communications with a simpler configuration. [Solution] The abnormality management device 1 includes a learning unit 12 that uses the characteristic directions of normal data as training data to learn, by maximum likelihood estimation, parameters of a probability model that outputs the posterior probability that the number of packets received in each time period corresponding to each of the characteristic directions of the normal data is normal; a derivation unit 13 that derives a probability distribution for the characteristic directions of abnormal data that indicates the number of abnormal packets received in each time period that deviates from the range of normal packet numbers, based on the posterior probability estimated by the probability model learned by the learning unit 12, the probability distribution for the characteristic directions of the normal data, and the prior probability of normality; and a calculation unit 14 that calculates a first index value that indicates the degree of spatial agreement formed by the probability distribution for the characteristic directions of the abnormal data derived by the derivation unit 13 and the probability distribution for the characteristic directions of the normal data.

A computer-implemented method and apparatus for unsupervised anomaly detection is provided. The method includes identifying one or more unsupervised anomaly detection approaches, wherein cach anomaly detection approach identified includes an anomaly detection algorithm and a corresponding threshold parameter value; and receiving a set of data. The method further includes, for the identified anomaly detection approaches, applying a statistical method including: sampling over the identified anomaly detection approaches to obtain a prior probability distribution; obtaining an input of anomalies and non-anomalies for at least a portion of the received set of data; obtaining a post probability distribution over the identified anomaly detection approaches based on the obtained input, wherein the post probability distribution updates the prior probability distribution; and determining whether a first stopping criterion is met and, if the first stopping criterion is not met, reapplying the statistical method. The method further includes recommending, based on the applied statistical method, one or more of the anomaly detection approaches. The method further includes for each of the one or more recommended anomaly detection approaches, applying a dynamic threshold optimization method including: comparing detected anomalies and non-anomalies with the obtained anomalies and non-anomalies; varying the corresponding threshold parameter value based on said comparison, to obtain an optimal threshold parameter value; and determining whether a second stopping criterion is met and, if the second stopping criterion is not met, reapplying the dynamic threshold optimization method. The method further includes identifying, for each of the one or more recommended anomaly detection approaches, the optimal threshold parameter value obtained. The apparatus includes processing circuitry and a memory containing instructions executable by the processing circuitry, whereby the apparatus is operative to perform the method for unsupervised anomaly detection.

The invention discloses an airport operation robustness decision-making method and device based on stochastic optimization. The method comprises the following steps: constructing a low-altitude intrusion situation feature space through multi-source heterogeneous sensing data; on the basis of the feature space, constructing a runway recovery time Wasserstein uncertainty set capable of dynamically adjusting the boundary so as to describe uncertainty; establishing a distribution robust optimization model for coupling aircraft position distribution and ground support facility scheduling; a column constraint generation algorithm is used for solving, and decision making and closed-loop feedback are executed in a rolling time domain mode. The method does not need to depend on prior probability distribution of interference events, adaptive balance of decision robustness and economical efficiency can be achieved, the problem of space-time mismatching is effectively avoided through resource coupling scheduling, and improvement of the operation recovery efficiency and safety of an airport under high uncertainty is facilitated.

The present invention relates to machine learning. To achieve communication-efficient distributed learning, a client is configured to obtain a prior probability distribution of parameters of a machine learning model in common with a server, and determine a codebook based on the common prior probability distribution. The client is further configured to obtain dataset and train the machine learning model using the dataset taking account of the prior probability distribution and the codebook. Through the training, an index of the codebook is obtained, which indicates a codeword in the codebook. The codeword is used as trained parameters of the machine learning model. The client sends the index to the server. In this way, communication overhead can be reduced, while precision of the learned machine learning model can be improved.

A method for generating a target function is provided. The method includes: performing normalization processing on a vector corresponding to each pixel in a target feature map set to generate a target vector, so as to obtain a target vector set; generating hash coding corresponding to each vector in the target vector set, to obtain a hash coding set; determining a prior probability of each hash coding in the hash coding set; and generating a target function based on an entropy of the prior probability.

The invention relates to a method and apparatus for single epoch position bounding. A method for determining a protection level for a position estimate using a single epoch of GNSS measurements, the method comprising: specifying a prior probability density of states x P(x); specifying a system model h(x) relating states x to measured observations z; quantifying a quality metric q associated with the measurements; specifying a non-Gaussian residual error probability density model f(r|θ,q) and fitting the model parameters θ using a set of experimental data; defining a posterior probability density P(x|z,q,θ); estimating states x; and calculating the protection level by integrating the posterior probability density P(x|z,q,θ) over states x.

The present invention discloses an LDPC decoding method, device, equipment and computer-readable storage medium, belonging to the field of decoding, and is used for LDPC decoding of data stored in NAND. When performing LDPC encoding and decoding, it is necessary to pad some data with zeros. Considering that during the decoding process, the data at the zero-padding positions in the variable nodes itself has a definite value and relatively high credibility. If a relatively complex decoding algorithm is used to update the prior probability and posterior probability of the zero-padding positions in the variable nodes, the decoding efficiency will inevitably be reduced. Therefore, in the decoding process of this application, a preset optimization update algorithm is used to update the prior probability and posterior probability of the zero-padding positions in the variable nodes. By setting a suitable preset optimization update algorithm, the amount of computation can be greatly reduced. On the one hand, the number of error corrections within a limited number of iterations can be increased, and on the other hand, the working efficiency is improved and the energy consumption is reduced.

The invention relates to the technical field of reservoir characterization, in particular to a reservoir characterization method based on a Viterbi algorithm. The method comprises the following steps: step 1, describing a spatial change mode of a prior probability of a reservoir facies by adopting a Markov chain, and constructing a reservoir parameter joint probability distribution model with longitudinal continuity; 2, determining a maximum posteriori estimation problem of reservoir parameters by combining the reservoir parameter joint probability distribution model according to a maximum posteriori estimation theory; and step 3, based on a Viterbi algorithm, developing and solving the problem in the step 2, and establishing a corresponding reservoir characterization model. According to the method, on the basis of the Viterbi algorithm, rapid estimation of multi-target parameters of the hidden Markov model is achieved, a brand-new reservoir characterization model is established for the actual problem of quantitative prediction of reservoir properties, longitudinal geological continuity is introduced, and synchronous prediction of multiple reservoir parameters is achieved.

The invention discloses a binary directional grey box fuzzy test method and system based on regional probability accessibility. The method comprises the following steps: S1, recovering a potentially missing indirect edge according to memory layout characteristics of a binary file; s2, obtaining a matching similarity score for each indirect edge, taking the matching similarity score as a corresponding prior probability, and calculating an accuracy probability of each indirect edge; s3, clustering into different regions according to the degree that the accessibility is influenced by the indirect edge, and constructing a region graph; s4, calculating the depth in the region and the connectivity between the regions according to the probability that each indirect edge is correctly recovered, and calculating the probabilistic reachability score of each path; and S5, performing optimal configuration on the fuzzy test according to the probabilistic reachability score of each path. According to the method, the directional fuzzy test based on reachability analysis under the binary condition can be effectively realized, the test strategy is adaptively adjusted when the control flow diagram dynamically changes, and the test efficiency and precision are balanced.

The invention discloses a method for realizing authorization verification for a deep learning model, and the method comprises the steps: a deep learning model owner carries out the extraction of an original model fingerprint through the input obtained through processing, and uploads the original model fingerprint and the identity information of the owner to a supervisor for binding and storage; establishing a prior probability statistical table based on the pirate model obtained by training and the collected irrelevant model; and then, iteratively performing pirate detection on the suspicion model based on the prior probability statistical table, calculating a confidence coefficient, communicating with a supervisor according to the confidence coefficient, and judging whether the suspicion model is a pirate model or not according to bound and stored information. According to the method, the problem of information loss caused by the fact that the pirate model only outputs the prediction label and does not give the corresponding probability can be solved, the pirate model can still be accurately judged under the condition that the output category of the model is changed, and the obtained detection conclusion is more reliable.

The invention provides a fault identification method and device and a program product, and relates to the technical field of data processing, the method comprises the following steps: obtaining at least one first node state parameter of a network node, the network node being a network node having a fault; calculating posterior probabilities of a plurality of fault types based on a conditional probability table and the at least one first node state parameter through a preset network model, the conditional probability table comprising prior probabilities and conditional probabilities between each fault type in the plurality of fault types and different node state parameters, the preset network model is used for calculating a posterior probability of a fault type according to a prior probability, and the plurality of fault types are a plurality of fault types when the network node fails; and determining a first fault type based on the posterior probability, wherein the first fault type is the fault type with the maximum posterior probability in the plurality of fault types. According to the invention, the fault identification accuracy can be improved.

The invention provides a communication method and device, and the method comprises the steps: determining rewards corresponding to K actions based on a reward model, and K is a positive integer; and sending first information to a first network device, the first information being used for indicating rewards corresponding to the K actions, and the first information being used for updating prior probabilities corresponding to the K actions. The first network device obtains which of the actions (or actions) can obtain a higher reward, so that the first network device can make a better decision. Moreover, the award is higher in accuracy compared with the prior probability, so that the decision making based on the updated probability is higher in effectiveness and reliability compared with the decision making based on the prior probability corresponding to each action.

The invention relates to the technical field of financial science and technology, and provides a claim assessment method, device, equipment and medium, and the method comprises the steps: obtaining multi-dimensional historical observation data of an insured object of a to-be-assessed insurance policy, and constructing historical observation features matched with the insurance policy type of the to-be-assessed insurance policy according to the multi-dimensional historical observation data; according to the historical observation characteristics, determining a prior probability that the to-be-evaluated insurance policy has a specified claim settlement event; obtaining multi-dimensional current observation data of the insured object, and constructing a current observation feature matched with the insurance policy type according to the multi-dimensional current observation data; and determining a target likelihood value of the current observation feature relative to the specified claim event, and correcting the prior probability according to the target likelihood value to obtain a posterior probability of the specified claim event of the to-be-evaluated insurance policy. The method can be applied to an insurance claim settlement evaluation task involved in the financial science and technology field, and the evaluation precision can be improved.

The application relates to the field of natural language processing, and discloses a risk text accurate semantic recognition method based on an improved deep learning model, which comprises the following steps: calculating the displayable width of a target segment by a character width counting rule and setting an upper limit, and establishing a trigger variable that automatically changes with a delivery condition; defining a display function under the width constraint to generate visible texts in different trigger states. Further, a trigger variable group is determined for each visible text, and a collision rate index is calculated based on the prior probability of the trigger variable, which is used to quantify the overlap degree of the multi-trigger mapping. Then, according to the index, masking sampling is performed after the trigger position to generate an input sequence, and segment representation is obtained through multiple sampling of the deep learning model; then, an adjustment parameter is calculated according to the collision rate index to adaptively adjust the temperature and noise of the model, and finally, a risk score is obtained through input of a classification layer.

This invention belongs to the field of wireless communication technology, specifically relating to a multi-user signal processing method for low-Earth orbit satellite communication. The method includes: acquiring a pre-compensated digital domain hybrid signal y; arranging the standard symbol sets of N users corresponding to y into a combined symbol set of N user combined signals s according to a preset user order; iteratively calculating the likelihood probability of all possible combinations of symbols corresponding to each element in y; and calculating, in the current iteration, the transmitted signal x of user n in the hybrid signal y based on all likelihood probabilities and the prior probability of each user updated through iterative feedback. n The posterior probability of all transmitted bits being 0 is calculated; for each user's posterior probability vector obtained in the current iteration, deinterleaving and channel decoding are performed; if the iteration terminates, a decision is made and the channel-decoded probability is output; otherwise, the channel-decoded probability is used as the new prior probability for that user, and the iteration is repeated. This invention ensures demodulation bit error rate performance.

The application relates to a causal DAG discovery method for an online service system based on fusion of soft priori. The method comprises the following steps: obtaining observation data and text meta-knowledge of the service system, preprocessing to form a standardized sample set, identifying variable types and semantics and outputting; generating a natural language description according to the variable semantics, querying a large language model for an ordered variable pair, analyzing to obtain three types of causal probability vectors, and calibrating to obtain an edge-level priori probability. An appropriate conditional independence test method is selected, high-confidence independent / dependent sentences are divided, and weights are assigned. A candidate directed acyclic graph is selected as an initial structure, parameters are estimated by linear regression, and data fitting scores are calculated, language priori scores, conditional independence penalty terms and counterfactual self-consistency penalty terms are calculated. Fusion is carried out into a hybrid score function, discrete optimization is carried out under the constraint of a directed acyclic graph, and a causal graph structure with the optimal score is output. The method can improve the efficiency and accuracy of a smart operation and maintenance system.

The invention discloses a search direction autonomous decision-making method fusing prior probability and a large language model, which belongs to the field of autonomous decision-making, and comprises the following steps of: firstly, aiming at a scene traversal problem caused by an unknown target position, performing scene search sequence decision-making through logical reasoning based on environment prior knowledge; generating an optimal search sequence as global guidance; and then, scene matching is realized through prior probability analysis and big language model reasoning based on visual observation, a region of interest is obtained based on semantic correlation evaluation of a big language model, and autonomous decision-making in the search direction is realized.

The present application relates to the technical field of data processing, and is a content data active exploration method and system based on cognitive graph and trace inference. The method comprises: encoding a multi-modal input signal of target content data into a trigger vector, and matching the trigger vector with a content cognitive ontology graph to anchor a target attribution node; automatically generating competitive hypotheses of each target attribution node in the content cognitive ontology graph, and constructing an exploration path; calculating the utility value according to the prior probability and the execution cost of each exploration path, and iteratively executing the exploration path in the order of utility value from large to small, updating the posterior probability of each competitive hypothesis in real time, and outputting an attribution diagnosis conclusion according to the posterior probability of each competitive hypothesis in response to the satisfaction of a preset termination condition. The present application can actively understand and verify the causality of content data, and improve the attribution accuracy of content data.

The invention provides a blasting vibration peak velocity prediction method fusing parameter uncertainty and data-driven optimization, which comprises the following steps of: firstly, acquiring data such as blasting parameters, lithologic indexes, geological conditions and actually measured vibration peak velocity (PPV), and establishing a basic database; a prior probability model is constructed, a joint uncertainty model is established in combination with probability disturbance and a fuzzy triangular number, a multi-dimensional disturbance sample is generated through a joint central value and a joint standard deviation and is fused with original data, and an extended database is formed. Feature analysis is carried out on the fused data, and input variables which have obvious influence on the vibration peak velocity (PPV) are screened; and constructing a BP neural network model based on the screened features, carrying out global optimization on a network weight and a threshold by adopting a PSO algorithm, and then carrying out local fine tuning by utilizing Adam. Finally, through training and verification, prediction of the vibration peak velocity (PPV) is realized, and model precision is evaluated through RMSE, MAE, MAPE, Rand other indexes. The influence of rock and soil parameter uncertainty on prediction precision can be effectively processed, and the reliability and applicability of blasting vibration prediction are improved.

The application discloses a device fault detection method and device and a storage medium. In one aspect, the prior probability of occurrence of each fault cause is determined based on the operation data of a screw compressor, and the likelihood probability of occurrence of a target fault under the condition of occurrence of each fault cause is determined, the posterior probability of occurrence of each fault cause corresponding to the occurrence of the target fault is calculated according to the prior probability and the likelihood probability, and the detection order of all fault causes when the target fault occurs is determined according to the size of the posterior probability of occurrence of each fault cause. In another aspect, the time sequence data of a characteristic parameter is monitored to determine whether the preset range under the current working condition is met, so that the reliability of subsequent fault causes is improved.

The invention relates to a segmented polynomial LDPC (Low Density Parity Check) decoding method, which comprises the following steps of: calculating a prior probability of channel output data, and generating a variable node matrix according to the prior probability of the channel output data in combination with a random construction method; calculating a check node expression through a complex function according to the variable node matrix, and performing approximation on the complex function by using a piecewise polynomial to obtain a check node matrix; and calculating the posterior probability of the channel output data according to the check node matrix, judging the posterior probability of the channel output data through soft judgment, if a check rule is met, completing decoding, otherwise, updating the variable node matrix again according to the check node matrix and the prior probability of the channel output data until decoding is completed. According to the method, the decoding performance and the implementation complexity are effectively balanced, and the problems that hardware implementation of a traditional BP algorithm is difficult, the complexity of a Log-BP algorithm is high and the performance loss of a Min-Sum algorithm is large are solved.

The invention provides a city POI distribution prediction method based on context learning, which belongs to the field of mobile communication and spatio-temporal data mining, and comprises the following steps: S1, inputting user signaling data and POI data into a mask model based on context learning to obtain a mask sequence; s2, using the mask sequence to train and finely adjust an urban POI distribution prediction encoder, extracting semantic information in the mask sequence, and obtaining a confidence probability that each region contains a certain type of POI; and S3, encoding data of known POI types contained in a small number of regions to obtain a mask sequence to be predicted, inputting the mask sequence to the trained urban POI distribution prediction encoder, obtaining a posterior probability in combination with the initial prior probability of the POI of the type and the confidence probability output by the model, and determining whether each region contains the POI of the type finally. The method does not need extra training or fine tuning, is suitable for different types of regional POI prediction tasks, and improves the prediction efficiency.

The invention relates to a system fault probability calculation method, which comprises the following steps of: acquiring fault modes of a system, and representing a system fault state as a binary vector containing all fault modes; for any fault mode, calculating the prior probability of the fault mode and the triggering probability of each detection node in the fault mode; based on the Bayesian theorem, according to the prior probability of each occurred fault mode contained in the system fault state and the triggering probability of each triggered detection node in each occurred fault mode, establishing a function relationship between the occurrence probability of the system fault state and the triggering state of each detection node; and obtaining the trigger state of each detection node, substituting the trigger state into the function relationship, and calculating to obtain the occurrence probability of the set system fault state.

The invention relates to the technical field of power transformation equipment, in particular to a power transformation equipment state evaluation probability graph network method and system considering operation conditions. The invention provides a power transformation equipment state evaluation probability graph network method and system considering operation conditions. The method comprises the following steps: acquiring original state parameters and operating environment data of the power transformation equipment; thirdly, establishing a state transition network, and training and solving an initial state and a prior probability of a model parameter through a deep learning algorithm; meanwhile, a working condition self-adaptive feature coding module is constructed to code the online state parameters to obtain a working condition prior probability. And finally, performing statistical evaluation by using a graph network reasoning algorithm, and outputting a result. According to the method, key problems in the prior art are effectively solved, and remarkable progress is made in the aspects of accuracy, adaptability, efficiency and interpretability. The method provides powerful technical support for improving the safety and reliability of the power grid, and has important theoretical value and wide application prospect.

The present application relates to an optimized expectation propagation detection method and a signal detection device. First, obtain the coding information, modulation information, channel matrix between the transmitting end and the receiving end, and received signal of the pre-coding bit sequence, construct a linear programming model, determine the prior probability of the transmitted symbol and the initial parameters according to the linear programming model, perform a first iterative convergence operation according to the prior probability of the transmitted symbol and the initial parameters, obtain the target log-likelihood ratio of the pre-coding bit sequence corresponding to the end of the preset first iterative convergence operation, and determine the pre-coding bit sequence according to the target log-likelihood ratio. In this method, more reasonable and effective initial parameters and prior information of the transmitted symbol are obtained for signal detection, so that compared with the traditional signal detection method in the present application, an accurate pre-coding bit sequence can be obtained with fewer iterative times.

The present application belongs to the technical field of tunnel and underground engineering geological detection, and discloses a method and device for sensing the lithology of a shield face, comprising the following steps: (1) selecting samples within a predetermined distance from a borehole in a sample point set to form a training sample set, and the remaining samples constituting a test sample set; (2) mapping the fluctuation scale of the vertical thickness of the lithology to an anisotropy scale factor in spatial interpolation, and then calculating the weight of each sample in the training sample set; (3) constructing an input feature vector; training a machine learning nonlinear model based on the input feature vector and a prior probability field to obtain a lithology real-time sensing model; (4) fusing the prior probability field and the sensing probability of the samples in the test sample set sensed by the lithology real-time sensing model to obtain a final lithology probability, and then reconstructing the shield face according to the final lithology probability. The present application improves the sensing accuracy.

The application discloses a method and system for representing sigmoid probability distribution based on a memristor, and relates to the technical field of the memristor, and comprises the following steps: receiving a sigmoid curve extracted from a memristor unit, and marking the sigmoid curve as a probability storage sigmoid curve; receiving input data, quantifying the input data into a probability value, and marking the input data as a prior probability; mapping the prior probability to a corresponding gate-end voltage value through the probability storage sigmoid curve, storing the gate-end voltage value to a node memristor array, and marking the gate-end voltage value as a storage probability value; quantifying the storage probability value to obtain a voltage signal, inputting the voltage signal to a weight memristor array, obtaining an output current, inputting the output current to a next node, and thus realizing a sigmoid belief network; and the application can directly represent the sigmoid probability distribution by using the current limiting effect of a 1T1R unit transistor, and simultaneously stores the probability value at the gate end.

This invention relates to a method, system, and equipment for establishing a pipeline risk analysis model based on fuzzy Bayesian networks, belonging to the field of pipeline risk analysis technology. The method for establishing the pipeline risk analysis model includes: constructing a pipeline failure analysis Bayesian network based on historical pipeline accident data and the Apriori algorithm; calculating the prior probabilities of basic risk factors based on expert opinions on basic pipeline risk factors and fuzzy comprehensive evaluation; obtaining the conditional probabilities of sub-nodes in the pipeline failure analysis Bayesian model based on expert pairwise evaluation of the relative importance of basic pipeline risk factors, as well as the analytic hierarchy process (AHP) and the ranking node method; and establishing the pipeline risk analysis model. This invention can calculate the failure probability of pipelines with complex multi-risk factors and identify key pipeline failure events. Based on the probabilities, it can assess the safety and reliability of pipeline systems, prevent accidents, reduce costs, and optimize resource allocation.

This invention provides an intelligent security risk prediction method and system based on a multimodal large model. The method includes the following steps: collecting video, audio, and sensor data, and outputting a unified event token sequence; generating natural language causal descriptions for the event token sequence, constructing an event causal graph with event tokens as nodes, causal descriptions as edges, and outputting logits as edge weights, and dynamically updating the graph structure; obtaining the prior probability of node risk using a Bayesian neural network, fusing the correlation information between nodes through a fusion graph neural network message passing mechanism, and using Do-Calculus intervention loss calibration to obtain the node risk probability; using a Continuous-Time Markov Network to infer future risk trends; outputting a decision report and executing a tiered response; and updating the end-to-end model under privacy protection through federated learning. The beneficial effects of this invention are: achieving cognitive-level risk identification and proactive intervention in complex scenarios; significantly reducing false alarm and false negative rates, and enabling early warning.

The invention provides an accident cause traceability early warning method and system based on an accident tree and a Bayesian network, and the method comprises the steps: constructing an accident tree model; converting the accident tree model into a Bayesian network model by using a predefined mapping relation; node prior probability: obtaining the prior probability of each leaf node, and obtaining the prior probability of each intermediate node based on a conditional probability table and the prior probability of each leaf node corresponding to each intermediate node, based on a conditional probability table and the occurrence probability of each root node corresponding to each intermediate node, obtaining the prior probability of the root node; priori early warning: performing early warning on an accident by utilizing the occurrence probability of a root node in the Bayesian network model; the method integrates the logic clearness of the accident tree and the bidirectional reasoning advantage of the Bayesian network, accurately locates the disaster source, quantifies the influence weight of each cause factor, and provides a scientific and effective disaster prevention, control and emergency management decision basis for a coal mine.