A data retrieval method, electronic device and computer-readable storage medium

Abstract

The embodiment of the present invention discloses a data retrieval method, electronic equipment, and a computer-readable storage medium. Starting from a node in a pre-established retrieval graph, the retrieval is performed in a predetermined retrieval manner to determine a set of candidate nodes and based on the candidate node The collection determines the retrieval results corresponding to the target data. In the embodiment of the present invention, when the similarity between the currently retrieved node and the target data is greater than the first threshold and the node matches the label information of the target data, the node is added to the candidate node set, and the currently retrieved node and the target data When the similarity of the node is greater than the first threshold and the label information of the node does not match the target data, the node is added to the candidate node set with a certain probability. Therefore, the embodiment of the present invention can support the retrieval of nodes that are similar to the target data and have matching label information. data, and can improve data retrieval efficiency.

Description

technical field [0001] The present invention relates to the field of computer technology, and more specifically, to a data retrieval method, electronic equipment and a computer-readable storage medium. Background technique [0002] A vector is a special data type consisting of several numeric values. In artificial intelligence and machine learning applications, vectors are widely used to represent features of data or entities. Vector search has become a common practice to find objects with the same characteristics as the search target. Taking picture search as an example, the system usually uses a specific method to extract the features of all pictures in the picture library to form a feature vector corresponding to the pictures one by one. When a user uploads a picture that needs to be searched, the system uses the same method to obtain the feature vector of the searched picture. Afterwards, among the feature vectors of all pictures, several vectors that are most similar...

AI Technical Summary

Problems solved by technology

With the rapid growth of modern application data, the number of vectors in the vector library is also increasing. It is gradually difficult to meet the needs of applications for vector search using traditional methods, thus promoting the research and application of various fast vector search methods.

Images