5results about How to "Meet search needs" patented technology

The invention belongs to the technical field of document retrieval, and particularly relates to a multi-modal long document retrieval method based on hierarchical representation and double-branch retrieval. According to the method, three-level hierarchical document representation composed of an element level, a page level and a chapter and section level is constructed by designing hierarchical features oriented to a document structure; a double-branch retrieval mechanism is used for conducting semantic-level matching and feature-level matching on queries, and a final evidence set is generated through a fusion strategy. According to the method, the cross-page semantic association and the fine-grained visual clues can be utilized at the same time, effective retrieval and information positioning of the multi-modal long document are achieved, the retrieval requirements of documents with different layout forms and different lengths in a real scene are met, and the method has important practical application value.

The application provides a time series data hybrid retrieval method, device and equipment for a security scene. The method comprises: data storage by using a multi-database cooperative storage architecture, the multi-database cooperative storage architecture comprising a vector database, a structured database, a cache database and a raw data backup database, each database storing corresponding type data according to a preset rule; according to a retrieval request of a user, retrieving results are obtained from the structured database, or a hybrid retrieval mode is started to obtain retrieval results from the vector database and / or the structured database; and the retrieval results are returned to the user. It can be seen that the application realizes efficient hybrid retrieval of structured attributes and high-dimensional feature vectors, improves the time series data retrieval performance and storage efficiency, and meets the large-scale and high real-time retrieval requirements of the security scene.

The invention discloses a method and a device for identifying a search intention of an automobile part, and relates to the technical field of computers. A specific embodiment of the method comprises the steps of performing entity recognition on an input automobile part search text to obtain a vehicle model entity, a part type entity and a specification parameter entity; semantic fusion is carried out on the vehicle type entity, the accessory type entity and the specification parameter entity and an adaptation rule in a knowledge graph, a fusion feature vector is generated, and the adaptation rule formed by the adaptation relation among the vehicle type, the accessory type and the specification parameter is stored in the knowledge graph; weighting the fusion feature vector based on the adaptation rule weight matrix; and processing the weighted fusion feature vector by adopting an intention recognition model based on an attention mechanism to obtain an intention recognition result. According to the embodiment, the recognition accuracy of the search intention is improved, and cross-vehicle-type fitting and rapid fitting of novel fittings can be realized.

The embodiment of the invention discloses a search component display method and device.According to the main technical scheme, the method comprises the steps that a playing page used for playing media files is displayed, a first search component is displayed on the playing page, and the first search component is used for skipping to a first search page after being triggered by a first operation gesture; the first search page is used for executing a search task; in response to the operation of triggering the component switching function, a second search component is displayed on the playing page, and the visual layout attribute of the second search component is obtained based on the behavior information of the user within the preset time and / or the basic information of the user. In this way, the visual effect is improved, the user can conveniently and rapidly search, and then the search experience of the user is improved.

This invention relates to the field of information retrieval and data visualization technology, and discloses a visual text search method based on a hybrid data model. The method includes: constructing a hypergraph hybrid model by associating a graph database and a vector database using global identifiers; identifying the intent of user queries; generating structured search instructions for the graph database and semantic similarity search instructions for the vector database; fusing and expanding the two types of search results based on global identifiers to form a search result set containing an extended knowledge graph, a list of semantic vectors, and traceable links between them; and finally, converting the result set into a dynamic and interactive visual view according to predefined mapping rules and a hybrid mechanics layout algorithm. This invention effectively unifies semantic retrieval and relational retrieval, improves the comprehensiveness and accuracy of retrieval, and significantly enhances the depth of information understanding and user experience through intelligent visualization and interaction mechanisms.