5 results about "General purpose technology" patented technology

The application provides a fault processing method and device and a storage medium, relates to the technical field of communication, and is used for solving the technical problem that general technology cannot quickly process network faults. The fault processing method comprises the following steps: acquiring a fault area of a fault network; acquiring operation information of a base station corresponding to a target operation subject in the fault area and not having a network fault; and when the operation information of the base station corresponding to the target operation subject meets preset bearing conditions, sharing network resources corresponding to the target operation subject by the base station corresponding to the target operation subject to user equipment in the fault area.

The application discloses a human resource vertical category model training method based on HR-RAG knowledge enhancement, belongs to the technical field of human resource vertical category model training, and comprises the following steps: step 1, knowledge structuring and building, step 2, designing a retrieval enhancement strategy, step 3, correlation coding and fusion, step 4, task alignment and incremental training, and step 5, confrontation verification and optimization. Through the HR vertical knowledge graph and the weighted retrieval strategy, the accurate adaptation of knowledge and business scenarios is realized, the general technical knowledge matching misplacement problem is solved, and the pertinence of model generated content is improved. Meanwhile, relying on the knowledge updating and incremental training mechanism, new rules and system iteration content can be quickly integrated, the knowledge lag problem is solved, and the output content can meet the latest requirements.

The application discloses a large model training method and system, electronic equipment and storage medium, the method comprises the following steps: determining the student model selection according to the domain attribute, quantitative performance requirement and resource constraint of the target task; determining the teacher model selection according to the thinking chain generation ability of the target task; inputting the to-be-processed data into the teacher model to obtain a distillation data set; fine-tuning the student model by the distillation data and the thinking chain data in the general technical scene to obtain a cold start student model; inputting the distillation data set into the cold start student model for multi-stage reinforcement learning training to obtain a target large model. Through the double screening of the thinking chain evaluation and the reasoning performance evaluation, the high-quality thinking chain generation ability of the teacher model and the efficient reasoning ability of the student model are ensured; the multi-stage reinforcement learning is trained in the gradient progression according to the thinking chain length, so that the thinking chain key step recall rate of the target large model is greater than or equal to 85%, and the reasoning performance is more than 4.4% higher than that of the traditional method.

The application provides a data management method and device and a storage medium, relates to the technical field of computers, and aims to solve the technical problem of low standardization degree caused by the difficulty of unified management of data by general technology. The method comprises the following steps: receiving a data collection request for requesting collection of initial business data; the data collection request comprises data source information, storage configuration information and metadata generation rules of the initial business data; the initial business data is mapped into target business data based on the data source information and the storage configuration information; and metadata information for describing the target business data is generated according to the metadata generation rules.

The application discloses a pipeline accident chain determination method, device, medium and product, relates to the oil and gas pipeline technical field, and aims to solve the problems of low efficiency and inaccurate analysis results of general technology in the aspect of oil and gas pipeline accident analysis. The pipeline accident chain determination method comprises the following steps: firstly, obtaining an accident text to be processed, wherein the accident text to be processed comprises text content of a pipeline accident. Then, a first accident element in the accident text to be processed and a co-occurrence frequency of at least two first accident elements are determined based on a graph neural network model. Then, a target co-occurrence network of the pipeline accident is constructed according to the first accident element and the co-occurrence frequency of the accident text to be processed. The target co-occurrence network is used to represent the co-occurrence relationship between the first accident elements, and the co-occurrence relationship is positively correlated with the co-occurrence frequency. Then, an accident chain of the pipeline accident is determined according to the target co-occurrence network.