241 results about "Best matching" patented technology

The invention relates to the technical field of retrieval enhancement generation, in particular to a method for automatically adjusting retrieval enhancement generation parameters based on content feature modeling. The method comprises the following steps: receiving an original query text of a user, performing component analysis, identifying terminologies, general vocabularies and question entities, and quantifying to form query fingerprints; acquiring a historical behavior sequence of the user, and constructing a score reflecting the level of the user by combining the query fingerprints and adopting a time decay weighting algorithm; the user level score is converted into specific retrieval parameter configuration, and a retrieval strategy blueprint is formed; guiding document library retrieval according to the retrieval strategy blueprint, and screening out a candidate knowledge set which is most matched with the professional level of the user; and according to the user level score, a preset instruction template is intelligently filled, and a situational generation instruction is constructed. According to the method, the problem of non-uniform cognitive load caused by a traditional system is solved through a retrieval enhancement generation technology, and the technical knowledge transmission efficiency and the user satisfaction are remarkably improved.

The invention belongs to the technical field of cloud computing and edge computing collaboration, and particularly discloses an intelligent scheduling and optimizing method for a cloud-edge collaboration network based on reinforcement learning. By constructing the state sensing matrix and generating the action decision vector, the problem that a traditional scheduling method is insufficient in correlation analysis of multi-dimensional operation state data in a complex network environment is solved, and the comprehensive sensing capability of the operation state of the network node is improved; a dynamic mapping mechanism among the running state, the resource limitation and the task allocation strategy is established, the task allocation and resource scheduling strategy is automatically and differentially adjusted according to the real-time state of the node, and the optimal matching between the task demand and the resource supply and the dynamic balance between the performance and the efficiency are realized; through performance index monitoring and closed-loop feedback optimization, the scheduling effect is mastered in real time, continuous iterative optimization is performed on the reinforcement learning model according to objective data, and resource waste and scheduling delay are reduced.

The invention discloses a heterogeneous computing system, a cache consistency maintenance method and device, equipment and a medium, and relates to the technical field of heterogeneous computing. The method comprises the step of redefining a cache coherence protocol of the heterogeneous computing system as a metadata index protocol, a multicast synchronization protocol and a hybrid response protocol. Entropy weights of different feature dimensions are determined according to feature values of all data blocks of the heterogeneous computing system, and a protocol most matched with the task load running state at the current moment is determined from a metadata index protocol, a multicast synchronization protocol and a mixed response protocol according to the entropy weights of the different feature dimensions; and determining whether to switch the protocol according to whether the current protocol is consistent with the selected protocol. According to the method, the problem that the protocol cannot be effectively optimized in related technologies can be solved, and the consistency protocol which is most matched with the current task load of the heterogeneous computing system can be accurately determined.

An example device for decoding video data includes a memory configured to store video data; and a processing system comprising one or more processors implemented in circuitry, the processing system being configured to: generate a prediction block for a current block of video data using a neural network; determine an equivalent intra mode for the prediction block from a set of available intra-prediction modes using decoder-side intra mode derivation (DIMD), the equivalent intra mode representing one of the intra-prediction modes that would generate an intra-prediction block that would best match the prediction block generated using the neural network; decode a residual block for the current block of the video data based on the equivalent intra mode; and combine the prediction block with the residual block to decode the current block of the video data.

The invention provides a well drilling and completion report self-adaptive SQL generation and intelligent question and answer data processing method. The method comprises the steps that S1, natural language query input of a user is received; s2, inputting user query input into a 'Zhongqi-7b-chat-1m' model, understanding a query intention, and outputting a most matched database table name; s3, accessing database metadata according to a table name output by the model, and obtaining structure information corresponding to the table; s4, inputting the natural language query and the obtained table structure information into a 'Zhongs-Code-32B-Instruct' model together, and generating an SQL (Structured Query Language) query statement conforming to semantic logic by the 'Zhongs-Code-32B-Instruct' model; and S5, executing the generated SQL query statement, and returning a structured query result. According to the method, the advanced natural language processing technology and the deep learning model are combined, so that automatic conversion from natural language query to standardized SQL is realized. According to the method, the query efficiency and accuracy are greatly improved, and the method is particularly suitable for non-technical personnel to perform complex drilling data query.

The invention provides a code reuse method and device based on AI drive, equipment and a medium. When a code submission event is detected, a code scanning process is triggered, a code analysis module driven by AI carries out multi-dimensional analysis on submitted codes, quantitative evaluation is carried out on the basis of a preset reusability evaluation model, and code snippets meeting a reuse standard are stored in a code knowledge base with a semantic index structure; analyzing annotation semantics based on a natural language processing technology through an AI retrieval engine, calculating semantic similarity between demand description and code snippets in a knowledge base in combination with a deep learning model, and positioning the code snippets which are optimally matched; after code multiplexing execution, performing incremental modification on the multiplexing code snippets and automatically generating change records; and carrying out value layering on the code snippets based on the Pareto analysis principle by using frequency trend prediction. Through combination of the AI technology and code reuse, the development efficiency can be remarkably improved.

The invention discloses a heterogeneous computing system, a cache consistency maintenance method and device, equipment and a medium, and relates to the technical field of heterogeneous computing. The method comprises the following steps: determining a hotspot data access path at the current moment according to an access sequence and dependency relationship strength of each data block of the heterogeneous computing system at the current moment; generating current running state information according to the hotspot data access path, the workload information of the heterogeneous computing system at the current moment, the cache effective information and the cross-device communication information; and based on the relationship between the running state of the heterogeneous computing system and the cache coherence protocol type, determining the cache coherence protocol most matched with the current running state of the heterogeneous computing system according to the current running state information. According to the method, the problem that the cache consistency protocol cannot be effectively and dynamically optimized in related technologies can be solved, the cross-device data dependency relationship can be dynamically captured, the dynamic change can be accurately sensed, and the high performance and data consistency of the system can be ensured.

The invention relates to the field of artificial intelligence, in particular to an intelligent diagnosis system for crop diseases and insect pests and a self-evolution method of the intelligent diagnosis system. The diagnosis system extracts features from a disease and insect pest image and a description text of a user for disease and insect pests, performs multi-modal fusion, generates a diagnosis report by using the fused features, and further comprises a gating network and a LoRA adapter which is specific for various disease and insect pest tags. And the gating network selects a corresponding LoRA adapter according to the description text of the plant diseases and insect pests, the corresponding LoRA adapter is dynamically injected into a Transform layer of the basic model, a reasoning path of the LoRA adapter is temporarily reconstructed, and the model with updated parameters is utilized to generate a diagnosis text. According to the method, the most matched expert capability can be loaded in time, so that the adaptability and the diagnosis accuracy of complex and changeable diagnosis scenes are remarkably improved.

The embodiment of the invention provides a CSI (Channel State Information) performance monitoring report triggering method based on an AI / ML (Artificial Intelligence / Markup Language) model and a related device, and provides a channel environment adaptive cumulative sum CUSUM algorithm, the algorithm configures different CUSUM parameter sets for different states of UE (User Equipment) respectively, each CUSUM parameter set comprises a relaxation factor k and a decision threshold h, and the relaxation factor k and the decision threshold h correspond to each other. The values of k and h in different parameter sets are different; the UE can select the most matched CUSUM parameter according to the current state (or scene), so that a performance monitoring mechanism of the UE can intelligently balance the detection sensitivity and robustness. Moreover, according to the scheme, performance differences are accumulated through a CUSUM algorithm and then compared with a decision threshold value to judge whether a performance degradation report is triggered or not, instantaneous jitter of a channel is effectively filtered through accumulation of the performance differences, and it is ensured that the report is triggered only when the performance of the AI / ML model continuously declines. Therefore, the uplink signaling overhead caused by mistaken touch and the resource waste caused by subsequent network response (such as mode switching) are reduced.

The invention provides a three-dimensional flow velocity field inversion method and system based on coherence-assisted SAR adaptive offset tracking, and a storage medium, so as to improve the problem of decoherence caused by time heterogeneity in a synthetic aperture radar interferometry (InSAR) technology, and relates to the technical field of electric power engineering geological disaster monitoring. The method comprises the following steps: establishing a space-time baseline network between SAR images; obtaining coherence data after geocoding; selecting a pair of SAR images from the space-time baseline network as a reference unit, calculating an offset result through a pixel offset tracking technology for the reference unit, and determining an optimal matching window parameter of the reference unit; establishing a mapping function between coherence and optimal matching window parameters; according to the established mapping function, calculating an optimal matching window parameter of each remaining SAR image pair; and in combination with the line-of-sight average deformation rate fields of the ascending and descending rails, inverting a surface three-dimensional flow velocity deformation field of the target area through a geometric vector decomposition model.

The invention relates to the field of data processing, in particular to an injection mold temperature control method and system, and the method comprises the steps: uniformly dividing the surfaces of injection molds of the same model and different batches into regions based on the area, and obtaining a historical temperature sequence and a real-time temperature sequence of each region; a sliding window is preset for the historical temperature sequence, the temperature fluctuation degree is calculated, and the temperature sensitivity degree of each area is determined. And calculating the similarity between the real-time temperature sequence and the historical production stage temperature sequence, selecting the historical stage with the maximum similarity as the best matching stage, and calculating the comprehensive temperature error. And according to the temperature sensitivity degree and the comprehensive temperature error, a temperature control signal of the PID control center is corrected, a comprehensive temperature control signal is obtained, and accurate temperature control is realized. By calculating the temperature sensitivity degree and dynamically adjusting the temperature control strategy, accurate and stable temperature control is achieved, manual intervention is reduced, and the production efficiency and the product quality are improved.

The invention discloses a storage goods allocation dynamic optimization method based on AI, and belongs to the technical field of intelligent storage and logistics management.The method comprises the steps that firstly, multi-dimensional data of goods, storage, business and the like are obtained; then, calculating a time-sensitive goods location-goods matching degree score integrating future delivery frequency of goods, association degree with adjacent goods, physical distance and a dynamic time urgency factor changing along with approaching of expected sorting time; a reinforcement learning model is used, the matching degree score is used as a key decision basis, a composite reward function considering overall layout gain, physical handling cost and high-urgency-degree task reward is used for guidance, and a dynamic goods allocation optimization scheme with globality and perspectiveness is output; according to the method, continuous self-adaptive adjustment of the storage layout can be realized, and the sorting efficiency and the response capability of the warehouse are remarkably improved.

The invention belongs to the technical field of artificial intelligence, and particularly discloses a retrieval enhancement generation method and system based on port equipment maintenance, and the method comprises the steps: obtaining equipment data according to a knowledge carrier of port equipment, carrying out the data preprocessing and data indexing of the equipment data, and storing the processed data in a vector database or an index; converting a fault query text input by a user into a high-dimensional real value vector by utilizing a pre-trained text embedding model; searching a fault knowledge fragment most matched with the high-dimensional real value vector in a vector database, and optimizing the fault knowledge fragment to obtain an optimal knowledge fragment set; and combining the optimal knowledge fragment set with the fault query text to generate an enhanced prompt, inputting the enhanced prompt into a large language model for reasoning analysis, and generating a fault positioning result and a processing suggestion. According to the invention, the accurate positioning of the fault mode can be realized, and the fault processing efficiency is improved.

The invention relates to the technical field of artificial intelligence and data engineering, in particular to an unstructured data universal structured acquisition method and system based on RAG. The method comprises the following steps: firstly, constructing a knowledge base for storing a table structure and field information of a service table; carrying out index identification and extraction on the input unstructured text by utilizing the large model; secondly, searching and matching a plurality of most similar business fields in a knowledge base aiming at each extracted index by utilizing a search enhancement generation technology; based on the input unstructured text and the most similar business field information of each index, assisting a large language model to elect a most matched business field of the index in a specific business scene; and finally, storing the index value of the index in a knowledge base in combination with the most matched business field. The universal data acquisition method for the unstructured data is suitable for index extraction universal for all service lines in the field, can automatically align and screen service fields in a specific service scene and store the service fields in the specific service scene, and realizes the universal data acquisition method for the unstructured data.

The invention relates to the technical field of conveyor regulation and control, in particular to a pipe chain conveyor operation regulation and control method suitable for bulk material conveying. Obtaining time sequence data of motor working states and monitoring point running states of the pipe chain conveyor in a preset time period and a historical time period, and screening out historical reference time most matched with working conditions at all moments in the preset time period by comparing difference characteristics of the motor working states in the historical time period and the preset time period; carrying out multi-dimensional operation state characteristic comparison on the material blocking probability and each moment in a preset time period, dynamically calculating the material blocking probability, and screening material blocking moments; the material blocking probability difference of all the monitoring points at the material blocking moment and the time relation of the material blocking moment are analyzed, an abnormal transmission chain is constructed, and a material blocking propagation path is revealed; on the basis, feature vectors are constructed, misjudgment is reduced in combination with a neural network, and a refreshed abnormal transmission chain is obtained; and finally, calculating the self-adaptive regulation and control intensity at the current moment based on the refreshed abnormal transmission chain to realize closed-loop optimization of a regulation and control strategy.

The invention provides a contract generation method and system based on artificial intelligence. Belongs to the technical field of artificial intelligence and comprises the steps that a user inputs basic information and personalized requirements of a contract through an interactive interface, and a management system analyzes the input information, extracts key features and constructs a requirement vector; constructing an intelligent decision model based on a reinforcement learning algorithm; a database containing various contract templates is constructed, and the management system selects the most matched template from the template database as a starting point of contract generation according to the output of the reinforcement learning model. Through an artificial intelligence technology, especially natural language processing and a reinforcement learning algorithm, user demands can be rapidly analyzed and contract texts meeting the requirements can be generated, so that the time of traditional manual contract writing is greatly shortened.

The invention belongs to the field of labor dispatch data analysis, and discloses a labor dispatch data analysis processing method and system and a storage medium. According to the invention, through dynamic matching and multi-scheme comparison of the personnel capability labels and the post demand labels, the optimal matching scheme can be accurately generated, and the labor dispatch matching success rate and the matching effect are improved. According to the method, the matching coefficient is generated by combining actual working data such as the efficiency coefficient and the stability coefficient, upgrading from static label matching to dynamic data driving matching is achieved, adaptability abnormity can be accurately recognized, the situation that the adaptability of matched personnel posts is insufficient is avoided, and the problems that the task completion rate is low, and client feedback is poor are prevented. According to the invention, a bidirectional correction system is constructed based on customer feedback and personnel feedback, and the mechanism can respond to business changes in real time and autonomously trigger a correction mechanism so as to quickly respond to actual business requirements, reduce manual intervention, greatly improve delivery management efficiency and reduce resource waste.

The invention provides a cross-platform WebView packaging device and method based on conditional compiling. The cross-platform WebView packaging device comprises a WebView packaging layer, a platform adaptation layer and a bottom platform layer, the WebView encapsulation layer provides a unified API interface as an entry directly called by a developer, and shields the difference of a bottom platform; the platform adaptation layer detects the type of a target operating system in a compiling stage by utilizing a Qt preprocessor macro, and dynamically selects and instantiates a bottom browser view most matched with the corresponding platform according to a detection result; and the bottom platform layer comprises a browser engine implementation class exclusive to each platform. According to the method, a unified packaging layer and an API interface are established, platform differences are digested at a bottom layer, and upper-layer service codes can be written once and run everywhere.

The invention relates to a computer system based on a computational model. A big language model-based computing device, method, computer readable storage medium and computer program product are disclosed. A computing device based on a large language model comprises a computing resource and the large language model, and the large language model comprises a decoding module used for obtaining a subsequent lexical element set by using original logarithmic score distribution; the retrieval module is used for retrieving the lexical element sequence to obtain a draft lexical element set; and the verification module is used for comparing the reference lexical element set with the draft lexical element set to obtain an output lexical element set. The computing device according to the invention overcomes the limitation of current retrieval-based speculative decoding in the aspect of matching normal forms, so that the most matched and most useful draft content can be efficiently and accurately found.

The invention belongs to the technical field of unmanned aerial vehicle multi-sensor fusion fault detection and elimination, and particularly discloses an unmanned aerial vehicle multi-sensor fusion fault detection and elimination method in a complex environment, and the method comprises the steps: constructing a visual feature matching data set, selecting an optimal model through hierarchical k-fold cross validation by employing a plurality of machine learning models, and carrying out the matching of feature points, error pairs are eliminated, an optimal matching set is found through an RANSAC algorithm, and visual input is eliminated when the effective point number is smaller than a threshold value; performing consistency analysis on the accelerated speed and the angular speed which are obtained through visual backstepping and IMU output, judging that an IMU fault occurs and adjusting the weight of the IMU fault when the threshold value is exceeded; iMU initialization is assisted by utilizing visual information; constructing real-time and accumulated residual chi-square test statistics for the GNSS, and setting a dynamic threshold value to detect a fault; traversing the satellite subsets to identify fault satellites and adjusting the weights of the fault satellites According to the multi-sensor fusion fault detection and elimination method for the unmanned aerial vehicle in the complex environment, the fault detection accuracy and the system robustness are improved, and safe and stable operation of the unmanned aerial vehicle is ensured.

The invention discloses a semantic tree index-based reasoning enhanced RAG retrieval method and system. The method comprises the following steps of: S1, reading and analyzing a document; s2, structuring the semantic knowledge tree by the document; s3, carrying out tree search and retrieval; and S4, enhancing the answer by the RAG system. According to the method, a long document of a specific field task is structured into an indexable semantic knowledge tree, and a large language model is guided to perform multi-step reasoning and path search on the tree in a retrieval process, so that a logic node which is most matched with query is dynamically locked, fundamental transformation from semantic similarity matching to structured logic positioning is realized, and the search efficiency is improved. The method does not need to depend on a vector database and embedded model training.

A system and method for online real-time multi-object tracking is disclosed. A particular embodiment can be configured to: receive image frame data from at least one camera associated with an autonomous vehicle; generate similarity data corresponding to a similarity between object data in a previous image frame compared with object detection results from a current image frame; use the similarity data to generate data association results corresponding to a best matching between the object data in the previous image frame and the object detection results from the current image frame; cause state transitions in finite state machines for each object according to the data association results; and provide as an output object tracking output data corresponding to the states of the finite state machines for each object.

Recommending machine learning models is provided. The method comprises training machine learning models, wherein each machine learning model is trained with a unique respective dataset. Metadata associated with each machine learning model is extracted, wherein the metadata includes properties of the respective dataset used to train the machine learning model. The machine learning models and metadata are stored in a model catalog. Upon receiving a new dataset, similarity scores are calculated between the new dataset and the machine learning models in the model catalog according to the properties of the datasets in the metadata of the machine learning models. A closest match machine learning model is identified from the model catalog for the new dataset according to similarity score. Responsive to a determination that the closest match machine learning model exceeds a similarity threshold, predictions for the new dataset are generated with the closest match machine learning model.

The invention belongs to the field of control, relates to a resin production waste gas recovery control method and system, and aims to solve the technical problems that in the prior art, working condition division cannot be matched with the real state of a multivariable system, and PID parameters are balanced among a plurality of control targets. The neurons are divided into stable working condition neurons, transition working condition neurons and disturbance working condition neurons; s2, obtaining a Pareto optimal solution set composed of a PID parameter group, wherein the Pareto optimal solution set comprises an ITAE optimal solution, a control output change rate optimal solution and an equilibrium solution; s3, determining an optimal matching unit of the potential feature vector in the self-organizing mapping network; s4, when the optimal matching unit is a stable working condition neuron and the absolute value of the current control error is greater than a high-order threshold value, selecting an ITAE optimal solution; when the absolute value of the current control error is smaller than a low-level threshold value, the optimal solution of the control output change rate is selected; and a balance solution is selected between the two. Accurate identification and division of the operation state of the resin production waste gas recovery process are realized.

A real time translation method for a game includes extracting features from video frames using computer vision and a database, performing an association process to find a machine learning model best matching the features for translation, obtaining texts in the game through optical character recognition (OCR), preprocessing the texts, translating the texts using the machine learning model to generate translated texts, and rendering the translated texts to images of the video frames for displaying the images with the translated texts on a display device.

Methods, systems, and computer storage media for providing security incident prioritization management using an adaptive incident prioritization engine of a security management system are described. The adaptive incident prioritization engine provides security incident prioritization based on an adaptive incident prioritization (AIP) framework built using a ranking algorithm. In particular, the adaptive incident prioritization framework employs a Best Matching 25 (BM25) algorithm and strategically and programmatically adapts the algorithm (e.g., an adaptive incident prioritization model) to rank security incidents based on a local security incident relevance metric (an adaptation of Term Frequency—TF—in BM25) and a global security incident rarity metric (an adaptation of Inverse Document Frequency—IDF—in BM25) associated with security incidents. A prioritization score for a security incident is calculated based on aggregating weighted frequencies of security incident ranking components (e.g., security incident metadata) within a security incident to determine an overall significance of the security incident.

The invention relates to an electric power document analysis and generation method and device based on a large language model and a medium, and the method calls the large language model to execute the following steps to perform document analysis and generation: obtaining a to-be-processed electric power document, and performing structured analysis and feature extraction to obtain first structured data; screening an optimal template which is most matched with the power document from a template library on the basis of the first structured data; and reading second structured data corresponding to the optimal template, verifying the second structured data, adjusting chapter titles according to the first structured data and synchronously updating writing cue words after verification is passed, outputting a document adjustment result, performing format standardization, and generating a final templated power document. Compared with the prior art, the method has the advantages of high power document template generation efficiency, high matching precision, self-adaptive adjustment capability and the like.

The invention relates to the technical field of medical image artificial intelligence, in particular to a multi-prototype cross-modal contrast learning fetal heart ultrasound image intelligent diagnosis system which comprises the steps that a to-be-detected fetal heart ultrasound image is obtained through an obtaining module, and a processing module inputs the to-be-detected fetal heart ultrasound image into a pre-trained image diagnosis model; the method comprises the following steps: performing feature extraction on a to-be-detected fetal heart ultrasound image, performing cross-modal multi-prototype comparison according to a feature extraction result to obtain a diagnosis category and a most matched image-text prototype pair, outputting the diagnosis category and the most matched image-text prototype pair by an output module, and finally performing CHD classification and screening according to the diagnosis category and the most matched image-text prototype pair. The problem that the model generalization ability is insufficient due to the fact that a single prototype center is mainly relied on to represent each category and the significant difference of the same category under different conditions is neglected is solved, and multi-category CHD intelligent screening is achieved by utilizing unlabeled images and a small amount of text priori on the premise that a large number of graph-text pairs are not needed.

The invention discloses a virtual human marketing method, system and device based on AI personality analysis, and the method comprises the following steps: collecting interactive dialogue data of a user and a virtual human in real time through a natural language processing technology, and extracting a multi-dimensional personality feature vector of the user through a deep learning model; according to the user personality feature vector, virtual human interaction parameters most matched with user personality communication are generated through a dynamic parameter adjustment model, and an emotional value is provided for the user; analyzing a short-term emotion demand vector of the user based on the real-time dialogue data; the system background constructs an emotion demand value vector matrix of the commodity library. Based on personality features and emotional features of the user, the virtual human can be close to the emotional demand of the user, and the use stickiness of the user to the system is improved; on the basis, the win-win situation of commercial value is achieved, virtual human parameter dynamic adjustment is added to enable the average dialogue duration of the user, more information of the user can be obtained, and the user portrait is more accurate.

The invention provides a full-scene adaptive intelligent recommendation method and system based on a graph neural network, and the method comprises the steps: constructing a parallel candidate recommendation model for generating a diversified candidate pool; a scene label of a user is determined through a CoT inference rule, the scene label is used as a query, a most matched strategy document is retrieved from a strategy knowledge base by utilizing an RAG module, and a currently recommended dynamic execution strategy is formulated; normalizing the original score of the parallel candidate recommendation model based on a dynamic execution strategy, calculating a preliminary fusion score according to a dynamic weight retrieved from a strategy knowledge base, and generating a preliminary sorting list; and performing rearrangement based on big language model reasoning on the preliminary sorting list in combination with scene and strategy guidance, and introducing evidence retrieved from a fact knowledge base based on an RAG module into a rearrangement result to obtain a final recommendation list with language interpretation. According to the method, the recommendation list which is highly personalized, self-adaptive in scene and clear in natural language interpretation can be obtained.