11results about How to "Fast retrieval" patented technology

The invention discloses an HNSW index algorithm optimization method and system based on a graph database and computer equipment, and the method comprises the steps: obtaining a current heat gradient table, comparing the current heat gradient table with a heat gradient table snapshot of a last optimization record, and obtaining a heat gradient table relative difference; only when the relative difference of the popularity gradient tables is greater than a set threshold value, optimization is carried out according to the following contents: data not in the current popularity gradient table are randomly sent to a low level of the HNSW index; inserting data in the current popularity gradient table into the upper hierarchy of the HNSW index, and updating the current popularity gradient table; after the optimization is completed, recording a current heat gradient table snapshot; by introducing the heat gradient table, the data hierarchy of the HNSW index algorithm is optimized, so that the retrieval speed of the high-frequency data is higher, the accuracy is higher, and the method is suitable for more business scenes.

The present invention relates to a scholar search method based on a cooperation network, and the method comprises the following steps: step one: constructing a scholar cooperation network; step two: optimizing the structure of the scholar cooperation network; step three: training a scholar node representation model; step four: establishing a scholar vector index; step five: searching scholars; the scheme realizes a search model based on word embedding, and compared with the traditional keyword search method, the search model based on word embedding can make full use of semantic information and improve the recall rate of the search.

The invention discloses a memory multi-thread fuzzy retrieval method. The method comprises the following steps: respectively calculating an editing distance similarity value and a common sequence similarity value of an entry and a user retrieval word by adopting a multi-thread parallel operation mode; and calculating a mean value of the editing distance similarity value and the common sequence similarity value corresponding to each entry and the search word, sorting the mean values, and selecting the first N character strings from the mean values as fuzzy search results. According to the method, under the condition of not depending on a big data retrieval component, a reliable result can be efficiently searched from million-level or even ten million-level entries in an extremely short time according to the retrieval word input by the user.

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 present application discloses a method for generating digital outline identifications of guqin score characters, wherein an input character string includes a character string representing a left-hand combination and / or a character string representing a right-hand combination; the input character string is identified to determine the left-hand combination or the right-hand combination; a connector is inserted between the codes of multiple display characters included in the left-hand combination and the right-hand combination to generate a digital outline identification sequence of the left-hand combination or the right-hand combination; a connector is inserted between the digital outline identification sequences of the left-hand combination and the right-hand combination to further combine the digital outline identification sequences of the left-hand combination and the right-hand combination to generate a digital outline identification sequence of score characters. The present application also includes a device for implementing the method. The present application solves the problem of large storage capacity and difficulty in expansion of score characters in the prior art.

A non-woven fabric for rapidly detecting organophosphorus pesticide residues comprises an enzyme-containing fabric, a water-soluble film, an indigo-containing fabric, a color-developing fabric and a flash-evaporation non-woven fabric, the enzyme-containing fabric and the indigo-containing fabric are respectively pure-cotton spunlace non-woven fabrics carrying cholinesterase and indophenol acetate, the color-developing fabric is a non-woven fabric, and the flash-evaporation non-woven fabric is a non-woven fabric. The top surface of the flash-evaporation-method non-woven fabric is connected with the bottom surfaces of the indigo-containing fabric and the color-developing fabric which are arranged side by side, the parts, except the indigo-containing fabric and the color-developing fabric, on the flash-evaporation-method non-woven fabric are connected with the four edges of the water-soluble film, and the top surface of the water-soluble film is connected with the bottom surface of the enzyme-containing fabric. The preparation method comprises the following steps: firstly, putting pure cotton spunlaced non-woven fabric into a cholinesterase gelatin solution for dipping and airing to obtain enzyme-containing cloth, then putting the pure cotton spunlaced non-woven fabric into an indophenol acetate solution for airing to obtain indigo-containing cloth, and finally compounding the cloth together according to a connecting structure between the cloth and the cloth to obtain the quick-test non-woven fabric. According to the design, the simplicity and the convenience are better, and the detection time is shorter.

The present invention discloses a neural network training method, a semantic similarity calculation method and a semantic retrieval system. The neural network training method comprises the following steps: S1: marking sentence similarity; S2: inputting two marked sentences into a neural network; S3: processing the features of the two sentences and calculating a loss function; S4: training the neural network according to the loss value of the loss function. The neural network training method used in the technical solution of the present invention has few neural network parameters, fast retrieval speed, can accurately calculate semantic similarity, and thus can obtain semantic information of the text, thereby improving the accuracy of retrieval, and is suitable for high concurrency and low latency demand scenarios.

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 invention relates to the technical field of computers, and discloses a retrieval method and device for recycled file information, computer equipment and a storage medium. The method is applied to first storage nodes. The method comprises the steps that when elastic search services are deployed in a first storage node and at least one second storage node, at least one retrieval condition from a management node is received, and at least one piece of first recycling file information is stored in the elastic search service of the first storage node; obtaining at least one piece of second recovery file information pre-stored in the elastic search service of each second storage node; based on the at least one retrieval condition, searching for target recycled file information corresponding to the at least one retrieval condition in the at least one piece of first recycled file information and the at least one piece of second recycled file information; and sending the target recovery file information to the management node, so that the management node feeds back the target recovery file information to the client. The problems that the retrieval efficiency of the recycled file is low and the user experience is poor can be solved.

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.