41 results about "Smart Cache" patented technology

A system and method for deep learning using a large codeword model with hierarchical caching is disclosed. The system processes input prompts into tokens, maps them to codewords using a codebook, and processes these through a machine learning core to generate responses. A sophisticated caching architecture stores and retrieves responses across both local and global cache tiers. The local cache maintains frequently accessed responses on edge devices through short-term and persistent storage components, while the global cache enables knowledge sharing across multiple devices. A context aggregator identifies relationships between cached responses to form comprehensive contextual representations. This hierarchical caching system significantly reduces computational requirements by reusing previously generated responses for similar prompts, while continuously optimizing cache contents based on relevance scoring and usage patterns. The approach enables efficient scaling across distributed environments while maintaining response quality.

The invention provides a real-time data distribution and intelligent caching method and system based on edge computing, and belongs to the technical field of computers, and the method comprises the steps: 1, collecting an original data stream of terminal equipment, converting the original data stream into a structured data format, and generating a feature vector based on the structured data; 2, outputting a first shunting signal and a second shunting signal through a pre-trained shunting decision model according to the feature vector; 3, extracting a feature identifier based on the first shunt signal, obtaining the feature identifier and a corresponding historical access record, and determining a final priority score of the corresponding data in the edge cache; 4, managing an edge cache space according to the priority score, and generating a cache update record; and 5, performing security processing on the structured data, generating a global cache distribution topology, and synchronizing the global cache distribution topology to an edge node. And redundant storage and cloud load are effectively reduced.

The invention relates to the technical field of artificial intelligence and automatic testing, and particularly discloses an AI-based automatic testing method and device. The method comprises the following steps: capturing and recording complete execution historical information in a process of executing a test task by an AI agent, and serializing and storing the complete execution historical information as reusable cache data; before the test is executed, the validity of the cache is judged through a multi-dimensional verification mechanism including data integrity, timeliness, functionality and the like; when the cache verification is passed, a result is quickly obtained based on the cache data playback test process; when cache verification is not passed or execution fails, triggering an AI agent to perform new reasoning and steps to generate and update cache data; and meanwhile, a caching strategy is dynamically adjusted according to test case changes and resource conditions. The device comprises a test executor, an AI agent manager, an intelligent cache manager and a dynamic cache management module, and all the modules work cooperatively to improve the test efficiency, reduce the calling cost of a large language model and enhance the stability of a test result. The method is suitable for Web / APP automatic testing, cross-platform testing, CI / CD process testing and other scenes, the execution efficiency can be remarkably improved, the cost is reduced, and resource utilization is optimized.

The invention provides an intelligent caching method and device based on a self-adaptive caching strategy, and the method comprises the steps: collecting an access event flow of each caching key in real time through a sensing probe embedded in a caching client, the event flow comprising Key, operation type, timestamp, request source, hit state and response time; running a streaming processing engine at a client based on an access event stream, and calculating indexes such as instantaneous QPS, continuous miss times and short-term access frequency by adopting a sliding time window; a calculation result is input into a rule engine for online matching, a decision instruction for a single Key is generated, and the decision instruction comprises distributed mutual exclusion lock addition, local TTL dynamic adjustment and read-write mode switching; a decision instruction is directly executed in a client memory through a strategy executor, and nanosecond strategy effectiveness is achieved; and finally, collecting system feedback data after strategy execution, inputting the system feedback data into a machine learning module for online learning, and dynamically optimizing a rule threshold value or recommending a new strategy.

The invention provides a multi-level distribution network path planning method and device, electronic equipment and a storage medium, and in the method, the total travel cost minimization under multiple scenes is taken as a target, and a resource redistribution mechanism and a cache acceleration technology are fused, so that the path planning efficiency is improved. The defect that adaptability, optimality and calculation efficiency of a solution cannot be considered in a traditional method is effectively overcome. Specifically, the method significantly improves the robustness of the final optimization path scheme for demand fluctuation, so that the final optimization path scheme can adapt to various actual scenes; meanwhile, through staged optimization and intelligent caching, the efficiency of the solving process is greatly improved, the consumption of computing resources is reduced, and the quality of a final solution is effectively improved, so that the final solution is closer to global optimum. Finally, the technical effect of quickly and automatically generating a high-quality and executable final optimization path scheme in a complex and uncertain environment is realized.

The application discloses a multi-model cooperative alarm log noise reduction and attack threat research and judgment method, and the specific process is as follows: firstly, the data preprocessing is carried out on the multi-source heterogeneous security alarm log, then the two-stage reasoning is carried out by using the cascade classification model based on the light gradient boosting machine, and the attack state is identified; for the high-priority alarm screened out, the context-containing prompt word is constructed, the large language model is called for deep research and judgment, and the intelligent cache mechanism based on the feature hash is introduced, so that the similar attack mode is quickly and accurately identified, and finally, the end-to-end security research and judgment report is output in combination with the automatic flow processing and breakpoint continuation mechanism. The application solves the problems of high alarm false alarm rate and low research and judgment efficiency in the current security operation.

The invention discloses an intelligent data service system based on a dynamic data source and an implementation method thereof. The system comprises a dynamic data source management module, an intelligent query engine, a filter chain processing system, a cache management system and an access point controller. The dynamic data source management module is responsible for dynamic loading, monitoring and management of data sources and supports dynamic switching and load balancing of the data sources. The intelligent query engine provides a unified query interface and supports query optimization of various data source types. The filter chain processing system realizes data processing and expansion before and after query, and supports dynamic script loading. The cache management system provides intelligent cache and performance optimization and supports dynamic generation of cache keys and flexible configuration of cache strategies, and the access point controller provides RESTfulAPI interface service and supports REST request processing of a client. According to the invention, dynamic management, intelligent routing and high-performance data query service of multiple data sources are realized.

The invention provides a distributed cache management method based on a dynamic strategy engine, and relates to the technical field of computer software, and the method comprises the steps: configuring a cache strategy for each tenant through an intelligent cache strategy configuration center, the cache strategy comprises a cache granularity, a cache updating opportunity, a cache life cycle, a cache replacement strategy, a cache updating mode and a cache medium; receiving a cache strategy configuration change through a cache strategy dynamic engine, and performing hot update of the cache strategy based on the tenant identifier; generating cache strategy recommendation based on system runtime indexes and a machine learning algorithm through an intelligent strategy recommendation system; caching strategy gray release is carried out through a multi-dimensional gray release manager based on the user ID range, the regional distribution, the equipment type, the access time and the multi-dimensional configuration of the service module; and the final consistency of the cache and the database is realized based on the message queue through the cache consistency guarantee module.

The invention provides a three-dimensional data rendering method and device, equipment and a medium, and relates to the technical field of three-dimensional data processing. The method comprises the following steps: determining a rendering node and a rendering instruction through a control node, determining whether target cache data associated with a to-be-rendered data identifier exists in an intelligent cache system or not by a target rendering node, and extracting the target cache data from the intelligent cache system under the condition that the target cache data exists in the intelligent cache system, and enabling the target rendering node to perform rendering processing based on the extracted data so as to generate an image or video stream of the three-dimensional scene corresponding to the user access request. Through real-time analysis of node load conditions, rendering tasks are dynamically allocated to corresponding display card resources, parallel processing of single tasks is realized, collaborative rendering of multiple tasks can also be realized, the rendering efficiency and the resource utilization rate are improved, an intelligent cache system is utilized to automatically adjust a cache strategy according to data access frequency, and the resource utilization rate is improved. And thus, quick access of common data is ensured.

The application provides a kind of based on streaming AI intelligent cache and breakpoint continuation method and system, the method includes: through the real-time encoding of the operation sequence of user and corresponding data to be uploaded by encoder, generates latent state sequence;The latent state sequence is input into the preset prediction model, and the prediction result is output, the prediction result includes: user next operation intention and affected data range;According to the latent state sequence, periodically generate and persist first breakpoint state snapshot;When breakpoint continuation is needed, the second breakpoint state snapshot selected by user from the first breakpoint state snapshot is loaded, and the replay operation sequence and replay data corresponding to the second breakpoint state snapshot are replayed based on the corresponding decoder of the encoder.

A PLC high-frequency data management system based on a built-in database comprises an embedded database, a database operation engine, a data recording function block, a database operation function block, an annular memory cache region, a nonvolatile storage medium and a communication service module during PLC operation. A lightweight database and an intelligent cache mechanism are deeply integrated in the PLC, and high-frequency real-time small data packets are written and converted into low-frequency batch large transaction operation, so that high-frequency, stable and structured recording and convenient query of process data are realized on the premise of ensuring control instantaneity and low CPU load.

Mechanisms are provided for caching threat intelligence enhancements in a threat intelligence cache of a threat intelligence computing system. The mechanisms, in response to a threat intelligence enhancement (TIE) being generated based on threat intelligence data, apply one or more threat intelligence cache policies to the TIE to determine a default cache retention period for a TIE type corresponding to the TIE. Different TIE types have different default cache retention periods. The mechanisms process the TIE and corresponding context data in the threat intelligence data to determine an adjustment to the default cache retention period based on the context data. The mechanisms modify the default cache retention period by the determined adjustment to generate a modified cache retention period. The mechanisms store the TIE in a TIE cache in accordance with the modified cache retention period.

The invention discloses a static single assignment form query architecture which solves the problems that an existing SSA analysis tool is low in query efficiency and large in resource consumption in a large-scale code library. The framework comprises a front-end SSA intermediate presentation layer, an intermediate cache layer and a background database layer, wherein the front-end layer converts a source code into an SSA code comprising a node, a data flow edge and a control flow edge; the middle cache layer adopts a dynamic self-adaptive size to cache an intra-process data flow sub-graph, an inter-process data flow abstract and a hotspot query result; and the background layer stores the complete SSA data structure chart through an SQLite database. Millisecond-level response is realized through intelligent cache, I / O and memory overhead are balanced by combining lazy query and lazy storage strategies, and static analysis efficiency and resource utilization rate of a large-scale code library are remarkably improved.

The invention discloses an intelligent cache processing method for real-time annotation data streams, and belongs to the technical field of cross-domain data processing, and the method comprises the steps: receiving a multi-source annotation data stream, mapping fields through a time sequence diagram neural network, calculating the similarity, generating a mapping rule, executing incremental cleaning, and outputting a structured annotation data stream; extracting data entities, attributes and relationships, constructing a heterogeneous cache atlas, and establishing a quaternary composite index; then constructing a heat-temperature-cold hierarchical cache architecture, dynamically distributing data entities based on real-time heat entropy, relation strength and the like, and synchronously updating the atlas and the index; low-confidence data is identified through entity confidence evaluation, correction suggestion feedback cleaning is generated, and data source subgraph copy parameters and hotspot entity states are synchronized by means of a security aggregation algorithm; evaluating the network based on a deep strategy to realize fine-grained dynamic access control; according to the method, the caching efficiency, the data credibility and the access security of the multi-source annotation data are improved, and the method is suitable for a real-time data stream processing scene.

The invention discloses a real-time big data stream processing method based on a multi-level cache mechanism, and relates to the technical field of big data processing and real-time computing, in particular to the real-time big data stream processing method based on the multi-level cache mechanism, and a system comprises a data access layer, a multi-level cache unit, a stream processing computing engine and a result output module. The core of the multi-level cache unit is that the multi-level cache unit adopts a three-level cache architecture which at least comprises a memory, a solid-state storage and a disk, and is provided with a cache management scheduler which is used for dynamically scheduling data blocks according to data access frequency and retaining hotspot data in a high-speed cache. According to the method, the system receives data streams through the data access layer, and the data streams are stored in the multi-level cache after being pre-classified; and the stream processing calculation engine preferentially obtains data from the cache for real-time calculation. Through intelligent cache management and data pre-classification, the data access speed and the storage capacity are effectively balanced, and the real-time performance of big data stream processing and the system throughput are remarkably improved.

The invention discloses an intelligent cache preheating method and system based on historical access data, and aims to solve the problems of insufficient prediction capability, low resource utilization rate and the like of a traditional cache mechanism. The method comprises the following steps: firstly, carrying out multi-dimensional data acquisition and distributed storage; extracting time, user, resource and system multi-dimensional features through data preprocessing and feature engineering; deeply analyzing and mining a time sequence rule, a user behavior cluster and a resource association rule through an access mode; then, a mixed prediction model fusing LSTM-ARIMA is constructed, and prediction precision is improved in combination with transfer learning and an online updating mechanism; a dynamic preheating strategy is generated based on a prediction result, and efficient preheating is achieved through high-probability resource screening, multi-objective optimization and batch scheduling; and finally, the effect is continuously improved through real-time monitoring, a fusing mechanism and closed-loop optimization. The method can improve the cache hit rate and reduce the system response time, and is suitable for distributed systems such as e-commerce and social networks.

The invention relates to the technical field of computer storage systems and high-performance computing, and particularly provides a storage optimization method and device based on double-buffer pipeline and intelligent cache scheduling, a four-level layered architecture of a main memory, a double-buffer pool, a high-speed cache and a computing unit is adopted, and all levels are connected through a bus; the main memory adopts a DRAM storage medium to store data; the double-buffer pool is composed of two independent local data buffer pools Buffer A and Buffer B, and both the two independent local data buffer pools Buffer A and Buffer B are realized by adopting an SRAM (Static Random Access Memory); the high-speed cache is divided into an L1 level and an L2 level which are both realized by adopting an SRAM; and the computing unit is responsible for executing target operation, directly reading data from the double-buffer pool or the cache, writing a computing result back to the double-buffer pool, and efficiently interacting with the storage hierarchy through an internal bus. Compared with the prior art, the efficiency of the storage system and the utilization rate of the computing unit can be improved, and the requirements of various memory access intensive computing scenes can be met.

A system and method for deep learning using a large codeword model with hierarchical caching is disclosed. The system processes input prompts into tokens, maps them to codewords using a codebook, and processes these through a machine learning core to generate responses. A sophisticated caching architecture stores and retrieves responses across both local and global cache tiers. The local cache maintains frequently accessed responses on edge devices through short-term and persistent storage components, while the global cache enables knowledge sharing across multiple devices. A context aggregator identifies relationships between cached responses to form comprehensive contextual representations. This hierarchical caching system significantly reduces computational requirements by reusing previously generated responses for similar prompts, while continuously optimizing cache contents based on relevance scoring and usage patterns. The approach enables efficient scaling across distributed environments while maintaining response quality.

The invention discloses a high-concurrency access method of a local password service interface and a multi-platform adaptation device, and solves the problems of poor high-concurrency performance and difficulty in cross-platform adaptation of a traditional scheme. The high-concurrency access end is used for constructing an API gateway through asynchronous non-blocking I / O, after protocol analysis and safety verification, a request is input to a four-level priority intelligent task queue, and expenditure is dynamically optimized in combination with a throughput model; tasks are distributed based on a dynamic load balancing mechanism of indexes such as CPU utilization rate and queue length, asynchronous processing and intelligent caching (the hit rate is larger than or equal to 85%) are matched, the throughput is improved by more than two times compared with a traditional scheme, and real-time task millisecond-level response is achieved. And the multi-platform adaptation end is used for generating a feature report through full-dimensional platform detection, matching an optimal algorithm and a hardware module by a self-adaptive execution engine, and unifying an interface to hide a bottom layer difference. Full-link protection and exception handling guarantee that the core service interruption rate is smaller than 0.005%, and the method is suitable for the key fields of finance, power grid dispatching and the like.

The application discloses a real-time intelligent cache processing method for labeled data stream, and belongs to the technical field of cross-domain data processing, comprising the following steps: receiving a multi-source labeled data stream, mapping fields and calculating similarity through a time sequence graph neural network, generating a mapping rule, performing incremental cleaning, and outputting a structured labeled data stream; extracting data entities, attributes and relationships, constructing a heterogeneous cache graph and establishing a four-element composite index; then constructing a hot-warm-cold layered cache architecture, dynamically distributing data entities based on real-time heat entropy, relationship strength and the like, synchronously updating the graph and the index; identifying low-confidence data through entity confidence evaluation, generating a correction suggestion, feeding back cleaning, and synchronously updating the parameters of each data source subgraph copy and the state of hot entities with the aid of a security aggregation algorithm; and realizing fine-grained dynamic access control based on a deep policy evaluation network; the method improves the cache efficiency, data reliability and access security of multi-source labeled data, and is suitable for real-time data stream processing scenarios.

The invention provides a data acquisition and real-time control system, which belongs to the technical field of data acquisition, realizes the integral jump of the system performance by constructing a closed-loop coordination mechanism among data acquisition, reliability analysis, real-time control and intelligent cache, and realizes the real-time control of the system by dynamically monitoring the load of an acquisition terminal and adaptively adjusting the acquisition precision according to the load. According to the method, high-quality acquisition of key data is guaranteed, overall resource consumption of the system is balanced, data congestion or resource idling is avoided, a key verification link is added for execution logic of the control system by taking a data reliability score generated in real time as a core basis of control decision, and the reliability of the system is improved. According to the method, limited cache resources can be preferentially allocated to data which is most critical and credible to system operation, a data storage structure is optimized, and the retrieval efficiency of key data during fault tracing and performance analysis is improved. Meanwhile, a closed loop of resource management is formed through feedback adjustment of a cache state on a front-end acquisition strategy, so that the system can intelligently perform self-adjustment when facing resource pressure, and continuous and stable operation of core functions is ensured.

The application provides an industrial software system data processing method based on intelligent caching, relates to the technical field of industrial software system data processing, and can create or activate a target cache partition corresponding to a new production plan by receiving a production plan change instruction, analyzing and obtaining business rule parameters of the new production plan, based on the parameters, preloading data related to a target production object from a back-end data source to the target cache partition according to a target vehicle model identifier, a target sub-component identifier and target sub-component version information, and routing a data request to the target cache partition for data searching when the data request from a workstation terminal is received. The application can dynamically adjust a caching strategy according to real-time production plan change instructions, create or activate a cache partition that is accurately matched with a new production plan, and pre-load related data.

The application discloses a kind of intelligent cache and prefetch collaborative scheduling method based on multi-dimension feature fusion.The method includes collecting the feature of multiple types of data, and preprocessing;According to the weight of the industry attribute setting multiple types of feature vectors, and calculate the fusion feature vector;Intelligent cache model is constructed, and fusion feature vector is input, and data access priority score is output;Intelligent prefetch model is constructed, and fusion feature vector and access time series data in history setting time are input, and future access probability and predicted access time are output;According to data access priority score and future access probability, calculate collaborative decision value, when collaborative decision value is greater than the first threshold value set, mark current data as priority cache data, and load priority cache data to cache at predicted access time.The application can improve cache hit rate and prefetch effectiveness, reduce data access delay, adapt to complex scenarios in multiple industries, and optimize cache space utilization.

The application provides a data collection method and system based on edge computing and intelligent cache strategy, and belongs to the field of Internet of Things data calculation. The technical scheme is as follows: collecting multi-source data and processing running conditions to generate unified data flow; constructing a cache object model and a system state vector based on the unified data flow, and implementing importance classification on the data of the cache object model; inputting the system state vector into a decision model to output a joint action, and correcting the feasibility of the joint action; executing the corrected joint action, processing the obtained data to realize hierarchical data supply; evaluating the effect of the hierarchical data supply to form an optimization closed loop. The application has the beneficial effects that: the application generates data flow in a unified time sequence on the edge side, constructs cache entries and importance classification, strengthens learning-driven prefetching, TTL and sampling counter-control, and guarantees A-class freshness and stable operation through rules, resource constraints and degradation fallback, thereby reducing backhaul and time delay and improving data supply reliability.

The embodiment of the invention relates to STM32 low power smart cache prefetching. In one embodiment, a method for operating an integrated circuit includes: sequentially requesting, by a processor of the integrated circuit, different instruction lines; when the processor processes a current instruction line provided in response to the corresponding request, determining, by a first comparator of the integrated circuit, whether at least one of the instructions in the current instruction line is a branch instruction by comparing the at least one of the instructions with a reference instruction; when the at least one instruction is a branch instruction, executing, by the processor, all instructions of the current instruction line prior to executing a next instruction line from a program memory of the integrated circuit; and when the at least one instruction is not a branch instruction, executing, by the processor, all instructions of the current instruction line before executing a next instruction line from the first volatile memory and the second volatile memory of the integrated circuit.

The invention provides an edge cache content predicting and caching method, device and system for the Internet of Vehicles. The method comprises the following steps: constructing a vehicle edge cache architecture; modeling a dynamic communication topological relation graph based on a real-time communication relation between nodes in the vehicle edge cache architecture; processing the dynamic communication topological relation graph by using a graph neural network to obtain a global feature vector; taking the global feature vector as attention bias, integrating the global feature vector into a prediction model based on Transform, and outputting future content popularity prediction distribution in combination with a historical request sequence; determining a hot content set according to future content popularity prediction distribution; and executing a collaborative content placement strategy among the macro base station, the road side unit and the vehicle node based on the hot content set. According to the method and the device, the multi-layer vehicle edge caching architecture is constructed and the dynamic graph neural network and the Transform model are fused, so that accurate prediction and intelligent caching of the content popularity in the Internet of Vehicles environment are realized.

The invention discloses an intelligent caching method and system for workpieces between automatic production lines and a production device, and belongs to the technical field of automatic production. The method comprises the steps that an MES system generates a warehousing task and issues the warehousing task to a WCS system, and the WCS system sends an instruction for transferring workpieces offline from a machining production line to a cache library to automatic transfer equipment; the MES system distributes a target goods location for the warehousing workpiece according to a preset storage location distribution rule; the storage location distribution rule comprises a constraint condition that only the same workpiece is stored in the same goods channel; when a preset ex-warehouse condition is met, the MES system selects ex-warehouse workpieces according to a preset priority rule and generates ex-warehouse tasks, the ex-warehouse tasks are issued to the WCS system, and the WCS system sends an instruction for transferring the workpieces in the cache library to a post-processing procedure to automatic transfer equipment; and the MES system preferentially combines the grouped workpieces to generate grouped ex-warehouse tasks during ex-warehouse according to the grouped processing requirements of the post-processing procedure, and matches the processing rhythm of the post-processing procedure.

The invention provides an adaptive fragmentation and cache collaborative optimization method and system for data distributed storage, and the method comprises the steps: constructing a space-time joint fragmentation model based on input data through time dimension processing and correlation quantification; adopting a cache prefetching engine to prefetch and cache the fragmented data of the spatio-temporal joint fragmentation model to obtain multi-level cache scheduling data; a distributed metadata management architecture is adopted, the fragmentation granularity of the spatio-temporal joint fragmentation model is adjusted, a caching strategy of multi-level caching scheduling data is generated, and self-adaptive fragmentation and caching collaborative optimization of data distributed storage is completed; according to the method, the designed intelligent cache prefetching engine is adopted, frequent access to a low-speed storage medium in the calculation process is reduced, and the cache prefetching efficiency is improved; a distributed metadata management architecture is adopted, adaptive fragmentation and cache collaborative optimization of data distributed storage is completed, and the problem of resource scheduling lag caused by independent operation of fragmentation and cache in a traditional scheme is solved.

Mechanisms are provided for caching threat intelligence enhancements in a threat intelligence cache of a threat intelligence computing system. The mechanisms, in response to a threat intelligence enhancement (TIE) being generated based on threat intelligence data, apply one or more threat intelligence cache policies to the TIE to determine a default cache retention period for a TIE type corresponding to the TIE. Different TIE types have different default cache retention periods. The mechanisms process the TIE and corresponding context data in the threat intelligence data to determine an adjustment to the default cache retention period based on the context data. The mechanisms modify the default cache retention period by the determined adjustment to generate a modified cache retention period. The mechanisms store the TIE in a TIE cache in accordance with the modified cache retention period.