46 results about "Application domain" patented technology

Systems and methods for constructing layered prompts to operate as input into a pre-trained large language model (LLM). The method involves obtaining a set of application domains in which the LLM will be used. Using these application domains, a set of guidelines is determined, defining operation boundaries for the LLM. A set of layers is determined, each associated with the guidelines and including variables representing attributes identified within those guidelines. Using these layers, a first layered prompt is constructed to test the initial operation boundaries of the guidelines and is supplied to the LLM to generate a set of responses. Based on the responses, a second layered prompt is dynamically constructed to test additional operation boundaries, ensuring iterative refinement and contextual relevance.

A decentralized artificial intelligence platform is disclosed that enables secure sharing and execution of AI modules in sensitive application domains. The platform comprises a knowledge module registry for discovering specialized AI modules, a security framework enforcing structural integrity of AI modules, a zero-knowledge proof validation service for verifying usage conditions without exposing private data, and a blockchain-based value chain for managing smart contracts and payments. In operation, an AI agent can fetch a required AI module from the knowledge module registry, prove the right to use it, execute the AI module on local data, and automatically pay the AI module's provider through a blockchain transaction.

The invention provides a hot updating method, system and equipment for an automation program of semiconductor equipment and a medium. The method comprises the following steps: dynamically loading a new version script code into an isolated runtime environment created by an application domain or a class loader through a reflective API (Application Program Interface) so as to isolate new and old codes. And performing preheating initialization on the new code in an isolated environment, wherein the preheating initialization comprises execution of an initialization function and simulation of a production request test. After preheating is completed, a new request of the production environment is routed to the new version to be processed through atomic switching operation, and meanwhile, the current running version continues to complete existing in-process affairs. According to the method, shutdown is avoided through dynamic loading and isolation, stability is improved through preheating initialization, new requests and in-process transactions are considered in atomic switching, non-shutdown updating of the automatic program of the semiconductor equipment is achieved, production interruption is effectively eliminated, and the equipment utilization rate and operation and maintenance flexibility are improved.

The invention provides an aerospace application knowledge graph construction method, device, system and equipment and a medium, and relates to the technical field of artificial intelligence, and the method comprises the steps: obtaining multi-modal data; based on the data types, data processing is carried out on the modal data, and lightweight markup language texts and embedded vectors are obtained; performing knowledge extraction on the lightweight markup language text on the basis of a preset cue word and a large language model to obtain metadata information, and constructing an entity relationship triple on the basis of a relationship between entities and the metadata information; and based on the aerospace knowledge file number information, determining an association relationship among the aerospace knowledge file, the lightweight markup language text, the embedded vector and the entity relationship triple, and constructing the aerospace application knowledge graph according to the association relationship and the entity relationship triple. According to the space application knowledge graph constructed by the method, the space application field entities and the association relationship can be accurately presented, and efficient and accurate knowledge support is provided for space application knowledge questions and answers.

Systems and methods for constructing layered prompts to operate as input into a pre-trained large language model (LLM). The method involves obtaining a set of application domains in which the LLM will be used. Using these application domains, a set of guidelines is determined, defining operation boundaries for the LLM. A set of layers is determined, each associated with the guidelines and including variables representing attributes identified within those guidelines. Using these layers, a first layered prompt is constructed to test the initial operation boundaries of the guidelines and is supplied to the LLM to generate a set of responses. Based on the responses, a second layered prompt is dynamically constructed to test additional operation boundaries, ensuring iterative refinement and contextual relevance.

The invention provides a feedback information generation method and device, and relates to the technical field of artificial intelligence, in particular to the technical fields of large models, automatic driving, intelligent traffic and the like. According to the specific implementation scheme, the method comprises the steps of determining an application field of current query information and a target component set associated with the application field; inputting the current query information and the target component set into a large language model, and generating a function component calling flow chart; and the execution function component calls the action instruction sequence corresponding to the flow chart to generate feedback information. According to the mode, the accuracy and efficiency of generating the feedback information are effectively improved.

The invention relates to a computer architecture for verifying image frame and image line data in a single pass. Aspects of this technical solution may increase the processing speed of low latency applications while maintaining the integrity of error correction detection at these higher speeds. For example, in an image processing environment associated with autonomous navigation (e.g., driving), a large amount of image data is to be quickly and accurately processed to maintain a reliable and latest model of the physical environment. Accordingly, embodiments according to the present disclosure may provide high-speed and accurate error correction of input frame data beyond the ability to be implemented by CPU processing or general GPU processing.

Embodiments herein describe a configurable engine integrated into a processor's cache hierarchy. The configurable engine can enable efficient data sharing between main memory, cache memory, and cores. The configurable engine can perform more efficient operations within the cache hierarchy. In one embodiment, the configurable engine is controlled (or configured) by software (e.g., an operating system (OS)) and is tailored to each application domain. That is, the OS can configure the engine according to the data flow profile of a particular application being executed by the processor.

Systems and methods for evaluating a pre-trained artificial intelligence (AI) model using layered inputs. The system obtains a set of application domains in which the AI model will be used, and a set of guidelines that define one or more operational boundaries of the AI model. The system determines a set of layers, where each layer is associated with corresponding guidelines and mapped to a set of variables and benchmarks. Each variable represents an attribute within the guidelines and each benchmark indicates the degree of satisfaction of the AI model with the guidelines. The AI model is dynamically evaluated against these benchmarks using a series of assessments. Subsequent assessments are dynamically constructed based on the outcomes of previous assessments. Scores are assigned to the AI model for each layer by comparing the expected and actual responses. The results are then displayed in a graphical user interface (GUI).

This invention discloses a multimodal event joint extraction method based on a link matrix. It determines the event structure system of the application domain and constructs a cross-modal event extraction dataset. A link matrix between image patches and text sequences is constructed. A multimodal event extraction model is built based on ViLT, where the multimodal Transformer encoder takes image patch pixel features, word instance features, and image patch-text word instance pair features as input to obtain a multimodal feature input decoder, acquiring image-text event type consistency features, event type features, and image-text event element features. Using maximum likelihood estimation, the image-text event type matching result with the highest probability, the event type, and the image-text event element are taken as an index, and the text event information is obtained by index mapping. This invention solves the problems of low recognition accuracy in weakly supervised and unsupervised methods and inaccurate matching of pipeline-style multimodal event element information, providing a new event extraction strategy for the field of multimodal information extraction.

The present invention provides a question-answer retrieval method, apparatus, device and storage medium. The question-answer retrieval method, through the comprehensive application of word segmentation sequence, feature sequence and mixed sequence of question text, firstly, integrates application domain knowledge and language knowledge into the structural analysis of question text, so that the calculation results can be more in line with the expectations of business personnel. Secondly, by utilizing feature sequences, the training corpus in the corpus can be deduplicated, reducing the impact of changes in the specific values ​​of features on the calculation results. On the one hand, the accuracy of the calculation is improved, and on the other hand, the requirements for the amount of change in the corpus in the corpus are reduced, reducing the workload and difficulty of corpus maintenance. Thirdly, by calculating and adding word order weights, the word order factor is added on the basis of cosine similarity, which improves the accuracy of the calculation and sorting results and enhances the overall retrieval effect.

Provided in the present disclosure are an intelligent multi-domain efficient collaborative information transmission method and apparatus, and a storage medium. The method comprises: acquiring N data packets generated by an application domain; determining configuration information of the N data packets in a control domain on the basis of an artificial intelligence technique and a multi-domain joint optimization indicator, wherein the multi-domain joint optimization indicator includes at least one of the following: a throughput, a transmission delay, a weighted transmission delay and a quality of service; and transmitting the configuration information of the N data packets to at least one communication node, such that the at least one communication node determines a transmission resource on the basis of the configuration information of the N data packets, wherein the transmission resource is configured to transmit the N data packets.

This invention discloses a cost- and reliability-aware profit optimization method for a heterogeneous multi-server system. The method includes: establishing a heterogeneous multi-server system model to obtain expected service revenue and multi-server system utilization; establishing a soft error reliability model to formulate the application domain investment selection and profit maximization problem as a constrained nonlinear optimization problem; utilizing an iterative method based on a penalty function to obtain the optimal multi-server system configuration, minimum cost, and maximum profit for any single application domain, and converting the profit and configuration optimization problem of the heterogeneous multi-server system into a 0-1 knapsack problem; and utilizing a dynamic programming-based algorithm and a greedy-based algorithm to solve the optimal application domain investment strategy and the corresponding maximum profit. This method can consider budget constraints and the soft error reliability of service requests while configuring the heterogeneous multi-server system configuration, thereby improving cloud service quality and provider profits.

The invention relates to a data synchronization method and device, electronic equipment and a storage medium, and the method comprises the steps: receiving a detection synchronization message for a target application in a multi-application cooperation system, sending a fault detection instruction to the target application, enabling the target application to carry out fault detection, and obtaining fault detection data; the detection synchronization message comprises a message propagation flag bit, a cross-domain flag bit and a to-be-synchronized application domain identifier; based on the message propagation flag bit, the cross-domain flag bit, the target address and / or the to-be-synchronized application domain identifier, determining a target synchronization application corresponding to the fault detection data from a multi-application cooperative system comprising a target detection application domain and a to-be-synchronized application domain; the fault detection data is sent to the target synchronization application, so that the target synchronization application executes the association operation based on the fault detection data, the message transmission range can be precisely controlled, the granularity and flexibility of communication are improved, and the fault detection efficiency is improved.

A data processing device (1901) for end-to-end user query correction and slot filling is described, the device having one or more processors for: receiving (1801) a user query (301); processing (1802) the user query to determine whether the user query is correct for an application domain language or a natural language; forming (1803) a corrected query (350) based on the processing of the user query; performing (1804) a slot fill operation on the corrected user query to identify one or more slots (306) of the corrected query, where each identified slot corresponds to a portion of the corrected user query; each identified slot of the corrected query is classified (1805) as one of a plurality of categories. In addition to the slot filling task, this approach can handle incorrect input variants in multi-language settings and can handle application domain languages in addition to different natural languages.

The application discloses a program processing method of a domain controller. The method is applied to the domain controller, the domain controller comprises at least two M cores, at least one A core, a first storage medium and a second storage medium, the first storage medium comprises a boot program of the domain controller, the second storage medium comprises M core application programs, a configuration file of the domain controller and A core related programs, and the method comprises the following steps: the A core checks whether each program in the second storage medium can be read and written; if each program in the second storage medium can be read and written, the A core updates to-be-updated programs in the second storage medium based on the configuration file and a downloaded update file; and if at least one program cannot be read and written, a first M core in the at least two M cores executes the boot program of the domain controller, formats the damaged program through the boot program of the domain controller, and updates the to-be-updated programs based on the configuration file and the downloaded update file.

A generator for converting an input vector from a latent space to one or more records x of measurement data that is realistic with respect to a given application domain. The generator includes: a trained neural network that is configured to map the input vector to a set of distribution parameters that characterize a random distribution of realistic measurement data, where this random distribution is configured such that given said set of distribution parameters and at least one source of randomness, samples of realistic measurement data may be obtained; and a sampling module including a random or pseudo-random number generator as a source of randomness and configured to sample the realistic measurement data from the random distribution.

The present disclosure relates to the field of network technologies and discloses an application access control method and system, and a device, a medium and a program product thereof. The present disclosure provides an application access control method. The method includes: generating, by a security management application client, an access request after detecting access to a target application, and sending the access request to a central domain name system; performing, by the central domain name system, domain name resolution on the access request to obtain a target application domain name of the target application, and sending the target application domain name to an application gateway; and determining, by the application gateway, a target access policy for the target application based on a matching result of the target application domain name in a first configuration file, and controlling, based on the target access policy, a terminal device.

The present application discloses a method, device, chip, and vehicle for upgrading the hardware domain of a chip. The method is applied to the chip and includes: if the first security domain determines that the chip enters the first upgrade mode, releasing the control of the first storage area of ​​the first security domain to the first application domain, where the first upgrade resource is stored; generating and storing first identification information used to indicate that the first upgrade resource has been released; in the first upgrade mode, the second security domain releases the control of its second storage area to the first application domain, where the second upgrade resource is stored; if the first application domain obtains the first upgrade resource and the second upgrade resource based on inter-core communication, performing an upgrade operation on the inactive partition of the first application domain based on the first upgrade resource, the second upgrade resource, and the third upgrade resource stored by the first application domain itself. This method implements the upgrade of a portion of the hardware domain.

The invention relates to the field of data processing, and particularly discloses a category activation method and device, electronic equipment and a storage medium. In response to a category activation request submitted by a user, a user domain module is called to obtain user information; in response to a first selection instruction for selecting at least one candidate account, calling an account domain module to obtain a category list corresponding to the selected target account; in response to a second selection instruction for at least one target category in the non-activated category list, calling the application field submitting module to perform preposition check on the at least one target category corresponding to the second selection instruction, and performing preposition operation when the at least one target category needs preposition operation; in response to submission information fed back by the application domain submission module, calling the user application domain module to activate and check at least one target category corresponding to the second selection instruction; and when the target category is activated and audited successfully, the activation result of the target category is fed back, so that the maintenance cost of the category activation function is reduced.

A local delivery service (LDS) operates in an enterprise computer network as a proxy for a communications server (CS) computer operating in a cloud computing environment. The LDS pulls a queue repository maintained by the CS and retrieves an output item flagged by the CS for local delivery. The output item has a document and associated configuration information specifying an application domain in the cloud computing environment and delivery settings specific to the application domain maintained by the CS. The LDS translates the delivery settings and routes the document to the connectors of communications channels of the enterprise computer network. In this way, documents produced by the CS in the cloud can be delivered locally to disparate destinations on the enterprise computer network without having to open individual ports in a firewall of the enterprise computer network to connect the CS with communications channels of the enterprise computer network.

This invention discloses a neural network reconfigurable configuration mapping method for FPGA resources, relating to the field of embedded artificial intelligence systems. The method includes the following steps: adjusting the FPGA platform according to different task requirements to construct multiple mapping schemes oriented towards resource constraints; establishing a mapping evaluation model by analyzing the resource characteristics of different FPGA platforms, evaluating multiple mapping schemes, and analyzing the differences between different mapping schemes and their performance differences on different FPGA platforms; and constructing a partially reconfigurable mapping adjustment scheme based on the deployment of different neural network algorithms according to the application domain. This invention analyzes the resource characteristics of the FPGA platform and effectively reduces the mapping space by imposing resource constraints. When a certain resource is limited, the accelerator performance boundary is further explored through the replacement of different resources with the same logic. It supports configuration mapping methods under the partially reconfigurable FPGA approach, and by analyzing the resource changes and parameter adjustments after partial reconfiguration of the target platform, a new mapping scheme can be quickly completed.

Embodiments of the present disclosure provide an intelligent multi-domain efficient collaborative information processing method, an apparatus, and a storage medium. The method comprises: acquiring N data packets generated by an application domain; on the basis of the N data packets, determining M feature sets corresponding to the N data packets, each feature set comprising at least one of the following features: a spectrum efficiency requirement, a transmission delay requirement, a reliability requirement, a capacity requirement, an energy efficiency requirement, and a quality of service requirement; in a control domain, on the basis of the M feature sets corresponding to the N data packets, determining M pieces of configuration information for the N data packets, each piece of configuration information comprising signal domain configuration information and resource domain configuration information; and, processing the N data packets on the basis of the M pieces of configuration information for the N data packets and intelligent technologies, N and M being positive integers.

An apparatus comprising a plurality of refinement nodes implemented on processors and memory, each arranged as a vertex in a graph topology. Each refinement node obtains inputs, performs refinement operations, and contributes enhanced outputs into a structured interchange maintained in memory accessible across the graph. The structured interchange is realized through embodiments including network-connected data containers, in-memory caches, or distributed key-value stores. Notification mechanisms signal availability of new or modified artifacts through publish-subscribe messaging, polling, processor interrupts, or equivalent signaling techniques. Refinement nodes opportunistically detect and act upon artifacts without centralized orchestration and without requiring synchronization of identical data copies. Refinement operations are cumulative and domain-specific, ensuring no node negates refinements of another. The apparatus operates as a computer-implemented collective system where successive outputs are progressively refined through iterative graph traversal, producing stable, convergent artifacts across diverse application domains.

Embodiments herein describe a configurable engine that is embedded into the cache hierarchy of a processor. The configurable engine can enable efficient data sharing between the main memory, cache memories, and the core. The configurable engine can perform operations that are more efficient to be done in the cache hierarchy. In one embodiment, the configurable engine is controlled (or configured) by software (e.g., the operating system (OS)), adapting to each application domain. That is, the OS can configure the engine according to a data flow profile of a particular application being executed by the processor.

The application relates to the computer technical field and provides an application filing result identification method and device, equipment and a storage medium. The method comprises the following steps: extracting application description information of a to-be-identified application and a plurality of access domain names involved in a running process from an installation package of the to-be-identified application; matching the plurality of access domain names with a preset third-party service domain name set, screening out each access domain name which does not hit any third-party service domain name as a candidate domain name; selecting a candidate domain name which meets a preset screening condition in the degree of association between the application description information of the to-be-identified application and each candidate domain name as an application domain name of the to-be-identified application; and matching the application domain name with prestored filed application data to obtain a filing result of the to-be-identified application. The above method can solve the problem that a server cannot accurately determine whether an application program has been filed.

This application provides a resource allocation method, apparatus, and device based on virtualization technology. The method includes: receiving a resource request sent by an application to be processed; in response to the resource request, obtaining virtual hardware resource information corresponding to the application to be processed, the virtual address corresponding to the application to be processed, and the virtual physical address corresponding to the application domain where the application to be processed resides from a configuration file; allocating corresponding computing resources to the application to be processed according to the virtual hardware resource information; and allocating corresponding memory resources to the application to be processed according to the virtual address corresponding to the application to be processed and the virtual physical address corresponding to the application domain where the application to be processed resides. The method of this application can support multi-domain fusion of the entire vehicle through a single controller, reducing computing power waste, the number of vehicle electronic control units, the length of vehicle wiring harnesses, and complexity.

The invention discloses an application traffic dynamic tuning method and device based on PBR and SR technologies, and the method comprises the steps: maintaining a key application domain name list according to user complaints and QoE monitoring conditions; periodically carrying out DNS (Domain Name Server) analysis on domain names in the key application domain name list to obtain a server IP (Internet Protocol) address list; dynamically generating a PBR strategy according to a DNS analysis result and issuing the PBR strategy; and through a PBR strategy, guiding the flow of the key application to a high-priority SR tunnel. According to the method and the device, a mature and simple technology is used, and flexible and effective key application guarantee is realized.

A system and method for processing multimodal data using jointly trained neural compression and enhancement networks. The system efficiently handles temporal, textual, sentiment, and structured data through specialized encoding modules. A fusion module integrates these encodings, capturing cross-modal relationships via attention mechanisms. The fused representation is compressed into a latent space by a trained compression network, then reconstructed and enhanced by a trained reconstruction network and neural enhancement network. Joint training optimizes all components simultaneously using a comprehensive loss function that balances reconstruction quality across modalities with enhancement performance. This approach enables superior data compression, reconstruction, and analysis by leveraging inter-modal correlations across various application domains. The system's latent space exploration capability facilitates generation of synthetic data scenarios for model testing and development.

This disclosure relates to a training dataset generation, dialogue processing method, device, medium, and product. The method includes: acquiring multiple first dialogue data samples from a target application domain; each first dialogue data sample includes input content and its corresponding response content; inputting the first dialogue data samples into a target generative model to obtain corresponding inference steps; the inference steps characterize the inference process of generating response content from the input content; integrating the first dialogue data samples and their corresponding inference steps to obtain a thought chain data sample corresponding to the first dialogue data sample; the thought chain data sample is data with a triple structure including sequentially arranged input content, inference steps, and response content; and constructing a training dataset corresponding to the target application domain based on each first dialogue data sample and each thought chain data sample. This improves the efficiency of acquiring thought chain data samples for the target application domain.