4results about How to "Fast retrieval" patented technology

This application relates to a large-scale model session method, apparatus, computer device, and storage medium. The method includes: utilizing a first search result generated in the previous search cycle based on a slow search method, and combining it with a second search result generated in the current search cycle based on a fast search method, to quickly determine a target search vector with high accuracy; using the high-accuracy first search result generated in the previous search cycle based on a slow search method, filtering out irrelevant noise in the second search result generated in the current search cycle based on a fast search method, to obtain a target search vector with high timeliness and accuracy. The target search vector is the vector corresponding to the memory search result, thereby achieving memory retrieval that is both reliable and accurate.

This invention presents an image retrieval method based on the fusion of attention enhancement and autoencoding. It uses an improved ResNet50 network to extract global and local feature maps, obtains global and local feature descriptors based on these maps, calculates image similarity, and derives the target image through similarity comparison. This image retrieval method improves upon traditional residual blocks by employing automaton encoding, effectively unifying local and global features into a single network. It uses an attention mechanism to extract regions of greater interest, avoiding algorithmic overhead and resulting in faster retrieval speed and higher accuracy.

The application provides a vehicle hardware image tracing method based on a control domain, comprising the following steps: S1, grouping and managing vehicle hardware, and establishing hardware coding sequence; S2, establishing and maintaining vehicle engineering EBOM; S3, establishing and maintaining vehicle manufacturing MBOM; S4, solving vehicle MBOM, obtaining vehicle controller information, and establishing a vehicle hardware list; S5, regularizing hardware information; S6, ascendingly sorting the regularized hardware, and calculating MD5 value of the sorting result; S7, comparing the newly calculated MD5 value with the existing MD5 value in the database; S8, version coding the MD5 value newly stored in the database according to the control domain; and S9, identifying whether the vehicle hardware information to be traced matches the vehicle hardware information in the database. The application provides a vehicle hardware image tracing method based on a control domain, which can quickly trace vehicle hardware information through vehicle hardware image, and the vehicle hardware image occupies small space, has fast retrieval speed, and has higher retrieval accuracy.

The application discloses a method for establishing a traditional Chinese medicine metabolite database, comprising the following steps: 1) establishing an LC-MS traditional Chinese medicine standard substance high-resolution mass spectrometry database containing retention time, first-order mass spectrometry information and second-order mass spectrometry information based on traditional Chinese medicine standard substances and a retention time correction internal standard group; 2) obtaining I and II phase metabolite samples of the traditional Chinese medicine standard substances through in-vitro incubation experiments, and extracting mass spectrometry information of the metabolites; 3) based on the traditional Chinese medicine standard substance database in step 1) and according to metabolic rules, a reaction type library, a reaction site library and in-vitro metabolism of the standard substances, metabolite information is summarized to form a metabolite database; and 4) combining the traditional Chinese medicine standard substance database and the metabolite database to obtain the traditional Chinese medicine metabolite database LutMet-TCM. The database provided by the application can meet the search requirements of traditional Chinese medicine compound prototype component, blood component and metabolite analysis and identification.