6 results about "Conditional probability" patented technology

The invention discloses a personalized aided teaching content dynamic generation and pushing method based on an AI large model, and belongs to the technical field of data processing, and the method comprises the following steps: collecting the conditional probability distribution of an output layer of the AI large model, and constructing a reference probability distribution vector; obtaining a generation probability distribution feature vector in a subsequent generation round and carrying out distribution similarity analysis to judge a generation probability distribution drift state; and performing controlled adjustment on the generated parameter set in a drift state, and constructing a stable window and dynamically updating the reference probability distribution vector in a continuous non-drift state. Teaching content is generated based on the adjusted generation parameter set and pushed in batches, generation stability monitoring is executed in the pushing process, pushing constraint adjustment is conducted according to the oscillation state and the convergence state, and finally the converged generation parameter set is stored as an initial generation strategy of the corresponding teaching task. The problem of recessive degradation of teaching content generation quality in the prior art is solved.

The application discloses a shipping accident analysis method, comprising the following steps: constructing a shipping accident knowledge graph; inputting a question, searching the shipping accident knowledge graph by using a search enhancement generation method, and generating an answer; wherein the search enhancement generation method comprises the following steps: analyzing the question by using a large language model, generating a query statement, searching the shipping accident knowledge graph according to the query statement, outputting nodes and edges matched with the query statement from the shipping accident knowledge graph, generating a Bayesian network according to the matched nodes and edges, generating a query result according to the conditional probability of a target node and its parent nodes and / or the posterior probability of the target node and its child nodes in the Bayesian network, inputting the query result into the large language model, and generating an answer according to the query result by the large language model. The application further discloses a shipping accident analysis device, a computer device and a storage medium. By using a large language model combined with a knowledge graph to extract and store shipping accident data, and by using a Bayesian network analysis model to analyze the shipping accident data, the occurrence probability of a complete cause-and-effect chain from causes to results of an accident can be obtained, so that objective and accurate causes and effects of a complex shipping accident can be sorted out, and reliable conclusion opinions can be provided for subsequent shipping business.

The invention discloses a power grid poisoning failure data recovery method and system based on attribute combination, and relates to the technical field of power grid data processing and information security, and the method comprises the steps: firstly marking a to-be-recovered data record and a failure attribute position after poisoning or failure data is detected; extracting various combinations of non-failure attributes in the record, and constructing a voucher group for estimating a failure value; counting all possible values and prior probabilities of failure attributes in the total data, and calculating the occurrence frequency of each voucher combination and the co-occurrence frequency of each voucher combination and candidate values; based on a conditional probability formula, the sum of the credibility of all the candidate values under all the voucher conditions is obtained, and the candidate value with the highest total credibility is selected to recover the failure position; and finally, writing the recovery value back to the data set to complete data recovery. High recovery accuracy can still be kept for high-failure-rate data, and key measurement data can be quickly recovered after a power system is subjected to data poisoning attack.

The invention discloses a generation power prediction method and device, electronic equipment and a storage medium, and belongs to the field of artificial intelligence. The method comprises the steps of obtaining background meteorological conditions of weather data prediction; inputting the background meteorological conditions into a trained conditional probability model to obtain conditional probability distribution output by the conditional probability model; disturbing the weather forecast data based on the conditional probability distribution to obtain disturbed weather forecast data; and inputting the disturbed weather forecast data into a preset power prediction model, and determining a power generation power prediction value according to an output result of the power prediction model. According to the embodiment of the invention, the conditional probability distribution of the weather forecast data space-time displacement error is predicted according to the background meteorological condition, the weather forecast data is disturbed based on the prediction result, and the power generation power is predicted according to the disturbed weather forecast data, so that the influence of the weather forecast data space-time displacement error on power generation power prediction is reduced; and the accuracy of power generation power prediction is improved.

The invention relates to the technical field of education intelligent evaluation systems, in particular to a civil education development evaluation method and system based on artificial intelligence, in the civil education development evaluation method and system based on artificial intelligence, through hierarchical aggregation calculation of a graph convolutional network, multi-dimensional features among nodes are dynamically integrated in a propagation path, the influence range of high-weight nodes is expanded, and the evaluation efficiency is improved. Low-weight nodes are replaced and reconstructed in an iteration process to form a resource linkage network with continuity and adaptive characteristics, a Bayesian network is adopted to carry out ratio reasoning on a resource proportion, resource adjustment is realized under budget and compilation constraint conditions, and through conditional probability updating of relatively high and relatively low nodes, the resource linkage network with continuity and adaptive characteristics is formed. According to the method, a distribution matrix with a self-balancing characteristic is established, so that resource input can be optimally adjusted according to a prior probability, node joint distribution calculation is used for capturing teacher stability and a course fluctuation trend, an observable risk area set is formed through statistical aggregation of fluctuation abnormal nodes, and an unstable structure is identified.

Embodiments of the present application provide a training method and an inference method of a flow matching generation model and related devices, which are used to improve the efficiency of the training process of the flow matching generation model. The method of the embodiments of the present application comprises: obtaining noise, a time t, a first modality object collected from the noise, and environmental features of the first modality object; inputting the noise, the time t, the first modality object, and the environmental features into an initialized flow matching generation model to obtain a predicted velocity field vector of a conditional probability path at a time t+1 output by the initialized flow matching generation model; calculating a loss between the predicted velocity field vector and a real velocity field vector by using a preset loss function, wherein the preset loss function comprises at least one of a first loss function and a second loss function, and a third loss function; and training the initialized flow matching generation model by using the loss and a back propagation algorithm until the flow matching generation model converges, to obtain a trained flow matching generation model.