1351 results about "On demand" patented technology

The invention provides an end-side multi-modal large model accelerated reasoning method and system, and the method comprises the steps: carrying out the two-stage screening and rearrangement of visual tokens based on the CLS attention and text-to-visual attention in a visual encoder and pre-filling stage, and constructing a sparse attention and sparse key value cache; in a decoding stage, an important neuron set is judged according to activation gating or historical statistics, only a corresponding feedforward network weight is pulled and calculated, missed weights are loaded on demand through asynchronous I / O, and hot neurons are maintained in a high-speed memory to utilize model sparsity, so that video memory / memory occupancy and calculation overhead are remarkably reduced on an end side; throughput and time delay performance are improved. According to the method, the internal memory and computing resources required by reasoning of the multi-modal large language model are reduced from two dimensions by utilizing the endogenous sparsity of the end-side large language model in input and the model, so that a higher reasoning speed is achieved by utilizing fewer resources on the premise of keeping the size of the model unchanged, and the performance of the whole system is improved.

The invention discloses a data processing system, method and device, a medium and a program product in the technical field of computers. In the application, the plurality of acceleration cards are directly connected with the plurality of memory devices through the management device, and the actual operation information of each acceleration card is collected through the management device, so that the memory space is allocated to each acceleration card as required according to the actual operation information of the acceleration cards. In other words, each memory device is not perceived by the host and is directly used as the extended memory of each acceleration card, so that the host does not need to distribute the memory for each acceleration card, thereby avoiding the condition of scrambling resources such as bandwidth easily occurring when multiple acceleration cards apply for the memory at the same time, and being not easy to cause data transmission delay. The purpose of expanding the memory for the accelerator card on the premise of not influencing the system operation performance is achieved.

The invention relates to the technical field of intelligent war game deduction, in particular to an intelligent war game deduction simulation system based on a digital twin fusion large model. Comprising a digital twin modeling unit which adopts a dynamic precision adaptive modeling mechanism; a large model decision unit; a deduction execution unit; and a dynamic interaction unit. A dynamic precision self-adaptive modeling mechanism is adopted by the digital twin modeling unit, the model precision can be dynamically switched on demand, the detail simulation demand and the computing resource consumption are balanced, and meanwhile, the sub-second-level dynamic synchronization of a virtual model and a war game deduction scene is realized through a time sequence calibration algorithm; and space confrontation, time sequence decision and rule triggering features are fused through the multi-modal feature analysis module, and the Wargame rule verification module is combined to embed entity performance boundaries and scene rule constraints, so that compliance confrontation decision parameters conforming to Wargame deduction logic can be generated, and the problem of insufficient model precision adaptation and decision constraints is solved.

The invention provides a multi-warehouse demand management method, device and equipment and a storage medium, and the method comprises the steps: carrying out the exchange rate sensitivity correlation analysis processing of demand fluctuation data in a cross-border e-commerce multi-warehouse network, and obtaining the exchange rate elasticity coefficient early warning information of each warehouse node; performing distributed coordination decision processing on the logistics constraint condition of each warehouse node according to the early warning information to obtain a decision scheme of resource allocation between warehouses; performing block chain trusted measurement processing on inventory distribution information in the multi-warehouse logistics alliance chain according to the decision scheme to obtain a multi-warehouse resource reconfiguration execution instruction based on the smart contract; and performing adaptive exchange rate risk avoidance processing on the inventory configuration strategy of each warehouse node according to the execution instruction to obtain a multi-warehouse collaborative inventory optimization management result. According to the method, the problem of influence of exchange rate fluctuation on multi-warehouse demand management is effectively solved through exchange rate sensitivity analysis and distributed coordination decision.

The invention provides a script preservation method which can be applied to the technical field of artificial intelligence. The script preservation method comprises the following steps: carrying out semantic understanding on demand related files by utilizing a large model, extracting core elements, and generating a structured demand change list; analyzing to obtain an abstract syntax tree of an existing script code, and constructing a mapping relationship between the software automation test element and a script code block; dynamically comparing the structured demand change list with an existing script code structure tree based on a mapping relation between software automation test elements and script code blocks, identifying a script range influenced by demand change, and outputting a change influence matrix; and based on the change influence matrix, generating an automatic script code snippet needing to be updated, and realizing preservation of the automatic script. The invention further provides a script preservation device and equipment, a storage medium and a program product.

Disclosed in the present invention is a general-purpose intelligent agent. The general-purpose intelligent agent comprises: an input module, which is configured to acquire a preprocessed input signal; a consciousness module, which is configured to perform perception and cognitive operations in a transparent and interpretable abstract thinking form; a self-awareness module, which is configured for perception and cognition of an intelligent agent body; a subconsciousness module, which is configured to use a data-driven model to implement perception and cognitive functions at a subconscious level; an information exchange module, which is configured to perform mutual exchange between perception, cognition and other information generated by the consciousness module, the self-awareness module and the subconsciousness module, thereby enabling integration and utilization; and an output module, which is configured to output a processing result on demand. In a machine intelligent agent of the present invention, a machine uses a general-purpose language as an underlying language for thinking and interaction, a thinking process and result are completely transparent, and thus autonomous decision-making, autonomous learning and continuous evolution are realized under fully controllable conditions. Further disclosed is a method for constructing a general-purpose intelligent agent and a general-purpose language. Compared with the prior art in which a natural language is used, the machine intelligent agent based on the general-purpose language has wide versatility in application.

The invention refers to method and devices for on demand paging for executing a sidelink relay path switch, the method including: Receiving a message from a remote user equipment, UE, including a measurement report and a list of candidate relay UEs; Selecting a target relay UE for the sidelink relay path switch from the list of candidate relay UEs; and performing one of: Transmitting a paging request to an entity of a core network, CN, including instructions to page or instruct the target relay UE to transit to an RRC_CONNECTED state for sidelink relay path switch purposes; and Transmitting a paging request to the target relay UE, including a request or instructions to transit to an RRC_CONNECTED state for sidelink relay path switch purposes.

The invention discloses a cross-film video stream non-inductive fusion playing method based on an HLS protocol, and belongs to the technical field of civil aircraft scene monitoring, and the method comprises the steps: 1, high-frame-rate screen recording preprocessing, 2, meta-information and video stream storage, 3, dynamic playlist construction, and 4, client non-inductive playing and rapid positioning. Through dynamic calculation of a playlist and a video frame index positioning technology, real-time loading and playing of any number of video clips with any time span, loading of target TS slices as required and asynchronous preloading of adjacent frames are supported, the smooth experience of zero lagging of cross-clip playing and comparable to real-time monitoring is realized, and a micro-batch preprocessing and on-demand loading mechanism is adopted, so that the real-time monitoring effect is improved. According to the method, the resource consumption of the server is reduced by more than 90%, the delay and resource bottleneck of a large-video multi-fragment scene are solved, the second-level event positioning capability of cross-day recording and broadcasting is provided for a civil aviation scene monitoring system, and the abnormal event analysis efficiency is greatly improved.

The invention relates to the technical field of software testing, and discloses a unit test case generation large model fine tuning method based on feedback reinforcement learning, which comprises the following steps: constructing a unit test fine tuning data set based on demand driving; an Actor model is constructed, and optimization is carried out through a demand-driven unit test fine tuning data set; a Critic model is constructed, and optimization is carried out through a demand-driven unit test fine tuning data set; and generating a unit test case based on the optimized Actor model, mapping the unit test case into a single instant reward based on the optimized Critic model, and obtaining a large model generated based on the unit test case through reinforcement learning. The unit test case generation method capable of automatically generating high-quality unit test cases with aligned requirements and consistent styles is constructed, the test efficiency and quality are remarkably improved, and the bottlenecks of traditional test generation in manual dependence, coverage depth and requirement consistency are broken through.

A wireless device may request system information, such as a system information block, from a base station on demand. The base station may indicate to the wireless device a change in system information via a message, for example, during a time period that the base station is not sending system information to the wireless device. The wireless device may expect to receive updated system information, for example, based on receiving the message, without having to send a request to the base station for system information.

The invention relates to the technical field of network operation and maintenance, and relates to an intelligent management and control method and device for an operation and maintenance work order, a computer program product and electronic equipment. The method comprises the steps of performing intention recognition on demand description information in an operation and maintenance work order based on a work order management and control agent, and determining a work order type and a work order task of the operation and maintenance work order according to an intention recognition result; according to the work order type and the work order task, generating an operation and maintenance guidance scheme of the operation and maintenance work order based on the operation and maintenance corpus knowledge base through the work order management and control agent; determining a target operation and maintenance worker based on the work order type, a work order adaptation model and a work order intelligent scheduling strategy through a work order management and control agent, wherein the work order adaptation model is a model representing the adaptation degree of the operation and maintenance worker and the operation and maintenance work order; and automatically distributing the operation and maintenance work order and the operation and maintenance guidance scheme to the target operation and maintenance personnel, and performing automatic order staring management and control on the operation and maintenance work order through the work order management and control agent so as to complete the disposal of the operation and maintenance work order. The management and control processing efficiency and quality of the operation and maintenance work order can be improved.

This invention introduces a sophisticated system for deploying projects on cloud platforms, combining Unikernels, Cyber Security Mesh Architecture (CSMA), MQTT protocol with SHA256 encryption, an Unikernel Orchestration Rules Engine (UORE), generative AI, and an innovative caching mechanism. Unikernels offer a secure, isolated environment for applications, reducing overhead and boosting performance. CSMA provides extensive security through analytics, identity management, and policy enforcement. The MQTT protocol, secured with SHA256, ensures the integrity and confidentiality of communications. UORE automates deployment, integrating a TLS terminator and data management for streamlined operation. Generative AI proactively resolves deployment challenges, particularly for complex applications, while the caching mechanism enhances performance and efficiency by minimizing latency. This integrated approach automates and secures the deployment process, enabling scalable, efficient, and real-time project creation and deployment in the cloud, thereby addressing the key challenges of cloud application hosting.

The invention provides a bill component design period and operation period micro-front-end extension method and device. The method comprises the following steps: developing a micro-front-end extension module according to a preset extension development specification; the micro-front-end extension modules are separated and packaged into different resource files according to different use scenes; registering the micro-front-end extension module as an operation period component or a design period component, and recording an access path and metadata of a corresponding resource file; when a receipt application is started, analyzing the configuration information of the receipt to obtain a core function and an extension point definition, and querying and identifying an extension list needing to be loaded by using registered extension description information and an access path of a resource file; and according to the extension list and the corresponding resource file path, dynamically loading the registered resource file of the running period component or the design period component as required. Through on-demand loading and dynamic module loading, a user only needs to load necessary extensions when opening a receipt, and the initial loading time is remarkably shortened.

Some embodiments provide order packaging control systems, comprising: a product packaging (PP) control circuit communicatively couple with an automated custom package fabrication (CPF) system configured to fabricate on demand custom packages, an automated product retrieval (PR) system configured to autonomously retrieve one or more products, pick stations, and a set of sensor systems. The PP control circuit can be configured to receive orders; prioritize the orders and define an order fulfillment sequence, track states of the PR system and the CPF system; determine product retrieval timing, and correspondingly determine fabrication timing of when the CPF system is to fabricate corresponding packaging; and control the PR system and the CPF system in controlling synchronous operation of the PR system and the CPF system to synchronize an availability of products of an order at a pick station and an availability of a custom package at the pick station.

The invention relates to a charging pile layout planning method and system based on demand analysis, and relates to the technical field of charging piles. The method comprises the following steps: determining a to-be-expanded area; determining the longitude and latitude of a charging pile station expansion point according to the position of the area to be expanded; determining the number of queuing vehicles in a crowded period; according to the historical charging information, obtaining the number of frequent charging users and the number of secondary charging users; according to the number of queuing vehicles, the number of frequently-charged users, the number of re-charged users and the historical charging information in the crowded period, calculating the total required charging time in the crowded period; and the number of the charging piles needing to be expanded is calculated according to the total charging duration in the crowding period and the number of the installed charging piles in the area needing to be expanded. According to the method, a to-be-expanded area is obtained through analysis, the total required charging duration in a crowding period is analyzed and determined by combining historical charging information of installed charging piles in the to-be-expanded area, the number of the to-be-expanded charging piles is designed according to the number of the installed charging piles and actual planning requirements, and the expansion points of the charging piles are determined by combining the position of the to-be-expanded area.

The invention provides a V2G service-oriented multi-granularity data on-demand sharing security aggregation method, and relates to the technical field of data security. The method comprises the following steps: acquiring multi-dimensional and multi-granularity power utilization data interacted between an electric vehicle and a power grid, and obtaining compressed data of each dimension based on super-incremental sequence compression; summing and aggregating the compressed data in the jurisdiction of the charging station, and obtaining an encrypted aggregation result by adopting Paillier homomorphic encryption; when a user accesses, a request is verified through an access authorization judgment function constructed by an RBAC model, and ciphertext aggregation (homomorphic addition is executed on global data, and differential ciphertext operation is executed on single-pile data to construct single-pile encrypted data) is executed as required after verification is passed; a complete decryption result is obtained through collaborative decryption by adopting a threshold decryption mechanism, and then target power consumption data is obtained through super-incremental sequence decoding. According to the method, on-demand access of multi-granularity data is realized, transmission overhead is reduced, data leakage and unauthorized access are completely eradicated through double guarantee of permission and encryption, and V2G multi-scene requirements are met.

The present application provides a graphical monitoring dynamic refresh method and device, which are applied to the field of information technology. By setting an interceptor in the persistence layer service, the change event of the database can be captured in real time. The change event is then pushed to the global Actor using the transaction message middleware. The global Actor is responsible for maintaining the full data view, and the local Actors filter out customized data on demand to meet the personalized monitoring needs of different users. Finally, the socket service is used to accurately push incremental data and only transmit the data that the user is concerned about, thereby significantly reducing unnecessary data transmission, greatly reducing the pressure on the database and network, and realizing millisecond-level customized dynamic refresh of monitoring data, which significantly improves the real-time performance and accuracy of the monitoring system and meets the strict requirements of terminal production operations for real-time monitoring.

The invention provides an intelligent speech recognition editor capable of efficiently processing large-scale texts. Belongs to the technical field of speech recognition and text editing. The editor comprises a voice recognition module which is used for receiving audio input, converting the audio input into a text and transmitting the text to a front end in real time through WebSocket / SSE; the text editing module is used for rendering and editing a text through a WebAssembly + virtualization technology; the data storage module is used for carrying out text caching by using IndexedDB / Local Storage; and the front-end UI interaction module is used for carrying out task distribution by adopting React / Vue in combination with WebWorker. According to the method, only the currently visible text area is rendered by adopting a block rendering technology and a virtualization technology, and the historical text is loaded as required in a lazy loading mode, so that the processing of million-level characters can be easily supported, and smooth operation and no blockage of an editor are ensured.

The invention discloses a self-adaptive hybrid power energy management method based on demand power prediction, and belongs to the field of hybrid electric vehicle energy management, and the method comprises the steps: 1, based on the historical information of a fixed route and the positioning information of a navigation system, utilizing an improved LSTM (Long-Short-Term Memory) to carry out power demand prediction on the basis of the historical information of the fixed route and the positioning information of the navigation system; carrying out certain time domain prediction on the vehicle speed and the required power of the fixed commuting vehicle through an LSTM (Long Short Term Memory) algorithm; 2, establishing an energy consumption minimum solving problem in a short time domain by estimating a required power range of a certain time domain, and solving an optimal basic equivalent factor sequence by using a particle swarm algorithm of adaptive weight; 3, equivalent factors under different required powers are corrected by considering the SOC and the temperature of the battery, an equivalent factor correction MAP is constructed through offline simulation and real vehicle calibration, and the equivalent factors are adjusted online in real time; 4, after the determined equivalent factors are obtained, an ECMS strategy is executed to determine the optimal output torque of the engine and the optimal output torque of the motor; according to the invention, the energy management control is close to global optimum, and the economic performance and working condition adaptability of the vehicle can be improved.

According to an embodiment of the present disclosure, a method, apparatus, device, and storage medium for information presentation are provided. The method comprises: presenting the target user's work information and a first category of user information from a plurality of categories of user information on the target user's information display page, wherein the category labels corresponding to the plurality of categories of user information are presented on the information display page; and in response to detecting a first switching operation for the user information, switching the first category of user information to a second category of user information on the information display page while maintaining the presentation of the work information. In this way, the presentation of user information can be made more reasonable, enabling users to quickly browse information on demand on the relevant information display page, thereby obtaining a better browsing experience.

The invention relates to the technical field of intelligent terminals, in particular to an intelligent terminal power consumption optimization method based on active and passive identification, which comprises the following steps: passive identification acquires a first signal data set by means of a low-power-consumption wireless communication unit and based on a low-power-consumption wake-up mechanism, active identification period and passive cooperative on-demand wake-up are carried out, and cooperative data acquisition is carried out; carrying out standardized preprocessing on the data, and carrying out weighted fusion on passive electromagnetic and active characteristics according to scene priorities to generate a comprehensive scene vector; querying a predefined strategy mapping table based on the current scene category, obtaining a strategy identifier uniquely associated with the category, and loading a corresponding strategy execution parameter set from a strategy library according to the strategy identifier; analyzing the strategy parameter, generating a time sequence instruction, executing clock gating on the target hardware unit in sequence, and sending instructions by the power supply and the bus arbiter to complete power consumption state switching. The accuracy of scene judgment is improved through an active and passive fusion mechanism, and power consumption regulation and control deviation caused by scene misjudgment is effectively avoided.

The invention relates to the technical field of mine equipment intelligent operation and maintenance and augmented reality, in particular to a forklift predictive maintenance and repair guidance system based on augmented reality, which comprises a multi-source data acquisition module, a mining area operation analysis module, a predictive maintenance engine, an AR intelligent repair guidance module and a resource dynamic scheduling center, a multi-source sensor is used for collecting equipment state and environment data, and mineral characteristics, path characteristics and load changes are dynamically analyzed in combination with working conditions of a mining area; the predictive maintenance engine constructs a fault prediction model based on a deep belief network, and dynamically corrects a working condition adaptive threshold; the AR module is used for positioning and superposing three-dimensional maintenance guidance through SLAM (Simultaneous Localization and Mapping), and adapting to the skill And the resource scheduling center generates spare part, personnel and alternative equipment scheduling schemes as required. According to the method, the whole process from fault prediction to maintenance execution is intelligentized, the fault early warning accuracy and the maintenance efficiency are improved, and the method is suitable for operation and maintenance management of mining forklifts under complex working conditions.

The invention discloses a cache maintenance system of a heterogeneous computing system and electronic equipment, and relates to the technical field of data processing, a first directory controller manages the consistency state of a host end, and a second directory controller manages the consistency state of an equipment end. The consistency maintenance operation is firstly efficiently processed by the corresponding directory controller in the storage domain, cross-device frequent coordination communication is reduced, and when cross-domain data access occurs, the two directory controllers cooperate through an established communication coupling mechanism and execute global consistency maintenance as required. The method not only can adapt to a high-burst and high-parallel access mode of a device end, but also can avoid direct conflicts with a cache management mechanism of a host end, and realizes global consistency through distributed collaboration. The technical problem of performance bottleneck of cache consistency in the heterogeneous system is solved, and the technical effect that the overall data access efficiency of the heterogeneous system is improved while the correctness is maintained is achieved.

The invention discloses a dye extraction device for textile detection and an extraction method thereof, and relates to the technical field of textile detection equipment. Comprising a supporting seat, an outer barrel is clamped at the top end of the supporting seat, a lower pressing seat for placing a textile sample is arranged at the bottom end in the outer barrel, a driving part is arranged on the outer barrel, an upper pressing seat is arranged on the driving part, and the driving part can drive the upper pressing seat to move downwards to extrude and rub the textile sample on the lower pressing seat; the squeezing table of the upper pressing seat is matched with the W-shaped containing groove, samples are evenly rubbed during rotation, it is ensured that dye is fully separated out, the reagent box and the balls are in linkage design, the reagent is released according to needs and evenly smeared, waste and manual intervention are avoided, the sealing cylinder automatically plugs or opens the drainage groove through the rubber sealing ring, and the liquid drainage time is controlled. The environment stability in the soaking stage is guaranteed, and the compatibility and the operation convenience of the device to different samples are enhanced in cooperation with a quickly assembled and disassembled medicament box structure.

Methods, systems, and computer program products for executing trusted software components in public computing clouds. Verifiably authentic software components are installed into a tenant partition of a multi-tenant public cloud-based environment. To do so, a software component installer is configured to install registered software components into the tenant partition. Installer processing includes (1) obtaining a component-specific token for a software component wherein the component-specific token is specific to both the software component to be installed and a particular tenant, and (2) installing the software component on behalf of the given tenant in the multi-tenant public cloud-based environment. Prior to executing the underlying code of the tenant-specific software component, the software component is authenticated with a component registry using the component-specific token. Additional trusted components are installed based on demand from within the tenant partition. No user or administrator intervention is needed and no credentials are hard-coded into the software components.

The invention relates to an inventory management method and system based on demand prediction. The method comprises the following steps: acquiring sales time sequence data, a point consumption curve, commodity shelf life data, external environment factors and other data; feature extraction is carried out on the integral consumption curve, and demand prediction processing is carried out based on sales time sequence data, integral urgency factors, regional integral density and external environment factors; calculating a safety stock threshold value of each warehouse according to the commodity demand prediction result and the integral urgency factor; and generating an optimal allocation path scheme based on the safe inventory threshold value and the current inventory state, generating a regional promotion scheme according to the commodity shelf life data, the meteorological data and the like, and executing the optimal allocation path scheme and the regional promotion scheme. According to the method, through demand prediction, inventory optimization allocation, promotion strategy collaboration and the like, the accuracy of inventory management is improved, complex demand fluctuation can be effectively dealt with, refined management and control of the inventory are realized, and the flexibility and adaptability of inventory management are enhanced.

The invention relates to the field of power systems, in particular to a load frequency control power system non-uniform sampling controller design method based on demand response. Comprising the following steps: establishing a multi-region load frequency control power system with demand response under non-uniform sampling control; constructing an integral quadratic constraint operator to quantify uncertainty caused by non-uniform sampling data and time delay; an original system is converted into a feedback interconnection system; a corresponding consistent index stability condition is given by using an integral quadratic constraint operator and a KYP lemma; designing a controller based on state feedback control and giving a corresponding iterative algorithm; the reliability and feasibility of the controller are verified through numerical simulation.

Systems and methods for customizing a product with on demand rendering. For example, embodiments of the present invention can determine dimensions of a physical space based on objects depicted in the image of the room. One or more processors can then generate a three dimensional virtual space based on the determined dimensions of the physical space depicted in the image of the room and generate a customized product based on selected customization options. The customized product can be rendered in the three dimensional virtual space based on the selected customization options.

The invention belongs to the technical field of database migration, and discloses a heterogeneous database migration scheduling method, equipment and device based on a distributed architecture, the method is applied to system services, and the method comprises the following steps: when a specified migration task is migrated from a database source end to a database target end, according to an object type of the specified migration task, scheduling the migration task according to the object type of the specified migration task; splitting the specified migration task into a plurality of migration stages with a dependency relationship; acquiring a preset migration sequence among a plurality of migration stages; and on the basis of the migration sequence, sequentially executing the corresponding migration stages by using the preloaded resources of the migration stages through preselected available migration service nodes to obtain execution results of the migration stages so as to complete migration of the specified migration task from the database source end to the database target end. According to the method, database migration in a complex migration scene can be realized, the migration success rate is improved, on-demand resource loading can be realized, and the migration efficiency is improved.

The invention relates to the technical field of power system optimization scheduling, in particular to a multi-region decentralized scheduling method based on demand flexibility aggregation, which comprises the following steps: (1) grouping and clustering transferable loads and interruptible loads to generate a flexible load aggregation model; (2) constructing an optimal scheduling model of the multi-region power system based on the flexible load aggregation model; and (3) decomposing each load cluster according to the optimal scheduling model of the multi-region power system, and ensuring that the decomposed load scheduling plan conforms to the user preference and system constraint of each load. According to the multi-region decentralized scheduling method based on demand flexibility aggregation, the scale of the scheduling problem is reduced through load aggregation, it is ensured that the scheduling result conforms to user preference and system constraints through load decomposition, the calculation complexity can be effectively reduced, and the renewable energy consumption capacity and economical efficiency of the system are improved.