42 results about "Service framework" patented technology

The invention relates to the technical field of intelligent operation and maintenance, and discloses an intelligent operation and maintenance method fusing multi-modal data and active learning, and the method comprises the steps: deploying a hierarchical Internet of Things equipment network in a target operation and maintenance region, collecting a heterogeneous operation and maintenance data set, and generating an operation and maintenance feature data set through employing a unified building operation and maintenance service framework cooperating with a plurality of MCPs; inputting the operation and maintenance feature data set and the user feedback information into a deep learning intention recognition model for intention recognition and demand analysis to obtain structured user demand data and an operation and maintenance task priority sequence; according to the operation and maintenance feature data set and the structured user demand data, performing real-time evaluation on the equipment operation state to obtain fault risk early warning data; autonomous decision analysis is carried out through an agent type artificial intelligence engine, an equipment maintenance scheme and a resource scheduling scheme are generated, and then the problems that in traditional operation and maintenance, fault prediction is single, the model adaptability is poor, and prediction results are difficult to convert into effective decisions are solved.

The invention relates to the technical field of model deployment, can be applied to business scenes of financial science and technology, medical health and the like, and discloses a model service cross-framework request conversion method, device, equipment and medium, and the method comprises the steps: receiving a reasoning request based on a first model service framework, analyzing input data and a request path, and extracting input parameters and model identification information; determining a target request path of the second model service framework according to the model identification information, and converting the input parameters into a target input format; sending the converted input data to a target reasoning service, receiving a reasoning result and extracting an output parameter; and converting the output parameter into an output format of the first model service framework and returning the output format to the client. According to the method, an adaptation mechanism is introduced between the client and the target model service, so that calling logic which originally depends on a specific reasoning framework can be accessed to heterogeneous reasoning service without changing, and structural compatibility and communication unification of cross-platform deployment and calling of the model service are realized.

The invention discloses a real-time translation recognition system under a cloud service framework, belongs to the technical field of real-time translation, and solves the problems that an existing translation system is insufficient in real-time performance, poor in scene adaptability, weak in privacy protection, slow in model evolution and the like. The dynamic model management engine obtains adaptive slices from a model slice factory according to scenes, equipment states and network quality and distributes the adaptive slices to edges, and the adaptive slices are distributed to a cloud-side collaborative reasoning system; the cloud-side collaborative reasoning system comprises a cloud-side collaborative reasoning system, a cloud-side collaborative reasoning system, a cloud-side collaborative reasoning system and a cloud-side collaborative reasoning system; the multi-modal perception engine fuses audio, images and dialogue history to generate a structured context vector and improve translation context fitting degree, the cloud edge cooperation engine takes an edge model as a core, processes different complexity tasks in combination with a cloud end, and constructs a data closed loop by incremental learning and a federation engine to realize model optimization and privacy protection; according to the system, the real-time performance, accuracy and safety are improved through cloud edge collaboration, dynamic adaptation and continuous learning, and the system is suitable for multi-scene real-time translation.

The invention provides a government affair question-answering method and product based on a knowledge-enhanced large language model, and relates to the technical field of artificial intelligence. A current government affair question-answering system mainly adopts a manual reply mode and is low in efficiency, and a system based on RAG lacks a query and discrimination mechanism and cannot recognize intention differences and utilize dialogue contexts, so that the quality of retrieval knowledge is influenced. Aiming at the defects existing in the related technologies, the invention provides a government affair question answering method based on a knowledge enhancement large model, and particularly, aiming at the problem of low manual reply efficiency, a complete government affair hotline service framework containing intelligent question shunting and structured feedback is designed; aiming at the problem of poor knowledge retrieval quality of a traditional RAG system, a multi-stage semantic enhancement retrieval method is designed, cross-domain knowledge is supplemented through networking query, and accurate response and service closed loop of government affair consultation are realized.

The invention relates to the technical field of communication, and provides a call chain monitoring method and a call chain monitoring system for browser requests. The method comprises the steps that a chain recording plug-in is called to obtain recording function starting operation and triggering time triggered by a user when a browser breaks down, and a monitoring identifier is generated; sending the monitoring identifier to a front-end request interceptor, and calling and starting a recording interface to execute recording; the front-end request interceptor adds the monitoring identifier into the received service request message and sends the service request message to a back-end service framework; the back-end service framework monitors the call chain of the message based on the monitoring identifier in the message to obtain call chain log data; the front-end request interceptor notifies a call chain recording plug-in to check recording data of the service request message; and the call chain recording plug-in queries the call chain log data. According to the method, the complete request calling link can be captured and displayed so as to identify the performance problem and the error source, and the troubleshooting time is shortened.

The invention discloses an AI-based micro-service system exception handling method and system and a storage medium, and belongs to the technical field of software exception monitoring. The method comprises the following steps: in an exception real-time capturing stage, deploying a global exception intercepting component at a micro-service gateway or each service framework layer, and generating an exception report; in the AI anomaly analysis stage, the anomaly report is analyzed, and a solution thought or a code repair suggestion is generated; an active alarm and notification stage: notifying development, operation and maintenance personnel according to an AI analysis result; meanwhile, all abnormal information is pushed to a monitoring panel; in the automatic repairing and deploying stage, exception handling is carried out, and dynamic deploying and system self-healing of exceptional patches are achieved; in the feedback learning and iterative optimization stage, exception processing results are collected, and the diagnosis accuracy and the repair suggestion quality are improved by continuously learning new exception cases and processing modes. According to the invention, abnormal information from different services is collected in a unified manner through a global interception mechanism, and the omission risk caused by decentralized processing is eliminated.

The present application relates to the field of intelligent operation and maintenance technology, and discloses an intelligent operation and maintenance method that integrates multimodal data and active learning. The method comprises: deploying a hierarchical Internet of Things device network in a target operation and maintenance area, collecting heterogeneous operation and maintenance data sets, and generating an operation and maintenance feature data set using a unified building operation and maintenance service framework coordinated by multiple MCPs; inputting the operation and maintenance feature data set and user feedback information into a deep learning intention recognition model for intent recognition and demand analysis, thereby obtaining structured user demand data and operation and maintenance task priority ranking; performing a real-time evaluation of the equipment operating status based on the operation and maintenance feature data set and the structured user demand data, and obtaining fault risk warning data; performing autonomous decision-making analysis through an agent-based artificial intelligence engine, and generating equipment maintenance plans and resource scheduling plans, thereby solving the problem in traditional operation and maintenance that a single fault prediction model has poor adaptability and prediction results are difficult to convert into effective decisions.

The invention provides a code migration method between software development systems and electronic equipment, and the method comprises the steps: determining a first communication data description file corresponding to a first software development system, converting the first communication data description file into a second communication data description file corresponding to a second software development system, and transmitting the second communication data description file to the second software development system, generating a service framework code according to the second communication data description file; determining a first business logic code corresponding to the first software development system, and converting first software interface information included in the first business logic code into second software interface information corresponding to a second software development system according to the second communication data description file, and generating a second business logic code corresponding to the second software development system and sending the second business logic code to the second software development system, so that the second software development system generates a target file executable by the target object according to the second business logic code and the service framework code. In this way, conversion and migration of software codes between software development systems are quicker and more convenient.

According to some embodiments of the present disclosure, the disclosure relates to an application system and server kit that create and serve digital twin-enabled applications. This disclosure also relates to a hub-and-spoke classification system. This disclosure also relates to a location-based services framework that leverages a generative content process to improve location prediction. This disclosure also relates to virtual reality and augmented reality applications, as well as digital agents that support various types of applications. This disclosure also relates to systems and methods for photometrically extracting information about objects and their features, including three-dimensional depth and related features.

The present disclosure provides a service implementation method and device based on a heterogeneous service framework, equipment and medium, wherein the application service implementation method based on the heterogeneous service framework comprises: accessing a probe on an application side, and adding a configuration file of the probe in a start script on the application side, the configuration file being pre-provided with an address of an application service of a data plane of the heterogeneous service framework, and the probe performing data interaction with the data plane based on the address of the application service. Through the embodiments of the present disclosure, the underlying code logic of the registration, discovery, calling, routing, flow control and the like of a plurality of service providers and consumers can be dynamically modified and reset to a unified implementation architecture, thereby achieving unified management of the heterogeneous service framework in a fully automated manner without invasion of the application development party, without introducing any new potential fault nodes, and improving the reliability and fusion of the heterogeneous service framework.

The invention relates to a model service operation method, computer equipment, a readable storage medium and a program product, and relates to the technical field of model services. The method comprises the following steps: in response to an operation instruction for target model service, obtaining a target model service template of the target model service from the operation instruction; matching the target model service template with each candidate model service framework to obtain a matching result between the target model service template and each candidate model service framework; screening out a target model service framework of which the matching result meets a matching condition from the plurality of candidate model service frameworks; and running the target model service framework based on the template content of the target model service template to obtain a running result of the target model service. By adopting the method, the operation efficiency of the template service can be improved.

The application discloses a method for constructing a dynamic and expandable single-agent runtime framework, and belongs to the technical field of artificial intelligence and large language model agents. The framework comprises a dynamic and expandable runtime framework module, a memory system module, a skill system module and a tool system module. The runtime framework module adopts a three-layer separation architecture and runtime pool management, supports LLM provider hot switching and a double-axis scheduling mechanism; the memory system module realizes five types of memory division, time-controllable context compression and experience sedimentation driven by a minimum subtask stage boundary; the skill system module realizes skill discovery through double-layer search, skill creation through an A / B evaluation cycle, and skill evolution through a closed loop using statistical tracking; and the tool system module adopts a four-level exposure model, a three-level permission control and a two-level dynamic routing mechanism. The application solves the problems of poor expandability and uncontrollable compression timing of existing frameworks, and realizes efficient, controllable and continuously evolving single-agent services.

The invention provides a micro-service platform management and control method for lossless management of different micro-service framework applications, belongs to the technical field of micro-service management and control, and aims to solve the problems that the difference between an existing Java micro-service framework operation mechanism such as Apache Dubbo, SpringCloud and HSF and a governance model is large, cross-framework interconnection and intercommunication are difficult, service governance is scattered, and codes need to be modified. A JavaAgent plug-in is loaded when a micro-service JVM is started, a framework core component is enhanced in a class loading stage, service registration, discovery and calling processes are intercepted, the framework core component is converted into a unified data model and a communication protocol, and cross-framework intercommunication and unified management are realized in combination with a unified management and control center comprising a registration center, a configuration center, a monitoring center and a service management center. According to the method, business codes and framework source codes do not need to be modified, lossless management of the micro-service is realized, and the management and control efficiency and the system stability are improved.

This invention discloses an automated tuning method and system for large language model service configurations, belonging to the fields of artificial intelligence and cloud computing technology. The invention includes an offline analysis module for performance profiling of the large language model (LLM) service system under a finite configuration set and for collecting training data; an online search module for exploring and evaluating the performance of different configurations in the configuration space based on a Monte Carlo tree search algorithm; and a simulator module for receiving the configurations generated by the online search module and predicting the performance metrics of the configurations under a target workload based on a logical step simulation model. This invention effectively solves the problems of low search efficiency and poor performance of existing automatic tuning methods in existing large language model service frameworks (such as vLLM, SGLang, etc.) due to the huge number of configuration parameters (reaching a petabyte-scale search space) and the complex dependencies and mutual exclusion relationships between parameters.

The present invention proposes a service-oriented loan repayment plan testing system and method. The system includes a Django Web service framework, a WEB UI module, an HTTP API module, and a repayment plan generator; the WEB UI module is used to provide a visual interface UI for users to access; the HTTP API module is used to provide external API interfaces; the Django Web service framework is used to provide URL mapping, HTTP protocol support, and MTV layered development support; the repayment plan generator includes a Views module, a loan parameter management module, a repayment plan module, and common tools. The present invention provides a loan repayment plan testing system that is easy to publish and maintain, and provides HTTP API interfaces for the integration of other programs or platforms, expanding the usage scenarios of the system.

The invention discloses a Nodejs-based progressive service framework, and relates to the technical field of computers, and the Nodejs-based progressive service framework comprises a double-layer architecture layer which is used for separating extended functions of core functions; the double-layer architecture layer comprises a core package and a framework template, the core package comprises operation architecture codes and general functions, the framework template appoints a system file structure and a function writing mode, the core package pre-integrates Modbus, SNMP and MQTT protocol analysis engines, and the plug-and-play equipment access capability is achieved through the standardized adaptation layer; an API (Application Program Interface) service framework and a communication management and data processing component are arranged in the core packet; the equipment access layer is used for automatically identifying the type of equipment when the equipment is dynamically or statically accessed, and carrying out conversion and access through the protocol adaptation layer; an operation and maintenance management layer; a modular design layer; an extension layer; according to the invention, a common equipment protocol analysis engine is preset in the core packet, and the equipment access process is simplified through the standardized adaptation layer, so that the complexity of development and integration is reduced.

The invention relates to the technical field of order management and control, and discloses an intelligent order dispatching method and system based on a concurrency mechanism language, and the method comprises the steps: obtaining the order information of a target order, and determining the distribution demand parameters of the target order according to the order information; the execution parameters of the target order are determined according to the allocation demand parameters and the order information, and the target order is dispatched according to the execution parameters, so that the interface service framework can be configured by utilizing the characteristics of high language compiling speed and high concurrent processing capability of a concurrent mechanism, the order determination efficiency is improved, the order processing preamble time is shortened, and the user experience is improved. Therefore, the instant inrush condition of massive orders in a high-concurrency scene can be effectively handled. And the order distribution demand parameters and the execution parameters are determined through the order information, so that the intelligence and refinement of order distribution are realized, the order distribution efficiency is improved, and the user experience is improved.

The invention discloses a service subscription method and device, a medium and a program product. The method relates to the field of distributed technologies, and comprises the following steps: when an application is started, generating a service list of the application through a micro-service framework; after an active subscription interface of the application is called to trigger subscription of the first service, obtaining subscription configuration information of the first service through the active subscription interface according to the service list; and subscribing the first service according to the subscription configuration information of the first service through the micro-service framework. By adopting the technical scheme, an active service subscription scheme is provided, and an active subscription interface can be provided for relatively important services in the application, so that the master control end can actively call the services as required, and the flexibility of service subscription in the micro-service architecture is improved.

The present invention relates to a dynamic task management method and system based on a service registration mechanism, belonging to the field of autonomous intelligent management of spacecraft; a finite state machine model of dynamic tasks to be executed in a task management module is established; the initial task state flag state is the idle state IDLE; the application layer initiates a registration request, and sets the task state flag state to the ready state READY; when the dynamic task start time startTime is reached, it becomes the occupied state BUSY; the specific service function registered at the dynamic task entry is started to be executed; when the dynamic task end time endTime is reached, it becomes the completed state DONE; the execution of the specific service function registered at the task entry is stopped; the application layer implements task cancellation, and the task state flag state of the task management module becomes the idle state IDLE; the present invention separates the task management service framework from the application layer, and the application layer can realize the wake-up and shutdown of tasks only by means of service request messages, effectively improving the flexibility of the architecture.

A deduplication service can be provided to a storage domain from a services framework that expands and contracts to both meet service demand and to conform to resource management of a compute domain. The deduplication service maintains a fingerprint database and reference count data in compute domain resources, but persists these into the storage domain for use in the case of a failure or interruption of the deduplication service in the compute domain. The deduplication service responds to service requests from the storage domain with indications of paths in a user namespace and whether or not a piece of data had a fingerprint match in the fingerprint database. The indication of a match guides the storage domain to either store the piece of data into the storage backend or to reference another piece of data. The deduplication service uses the fingerprints to define paths for corresponding pieces of data.

The present application relates to the field of computing and processing technology, and provides a log retrieval method, device, electronic device and storage medium, the method comprising: respectively initializing the classes on the service framework of multiple clients to obtain the unique identification code corresponding to each client; respectively obtaining at least one log printing interface of each client according to the log printing signal, and obtaining the subclass of the log printing interface; printing the log of the log printing interface through the subclass of the log printing interface to obtain multiple logs with corresponding unique identification codes; and retrieving the log from multiple logs with corresponding unique identification codes through the unique identification code corresponding to each client to obtain the target log. The technical solution of the embodiment of the present application can print out a log with a unique identification code, and can quickly and accurately retrieve the target log from the full-link log through the unique identification code.

In general, certain embodiments described herein relate to a method for performing inferences on noisy compressed data. The method includes receiving, from a client, the noisy compressed data at a machine learning system. The noisy compressed data is transmitted over a network with the machine learning system being a machine learning as a service configuration. An amount of noise in the noisy compressed data is based on a privacy level. The method further includes classifying, by a classifier of the machine learning system, the noisy compressed data by executing an inference on the noisy compressed data to obtain classification results, and transmitting the classification results to the client via a network. The inference is executed by a trained model, which is trained on noisy compressed training data. By transmitting noisy compressed data, the privacy of actual data is preserved.

The present application provides a target software upgrade method and system, computer equipment, and readable storage medium, including: generating an unordered block mirror file system and a compressed root file system based on the page size and sector size of the storage medium, respectively, and packaging the unordered block mirror file system and the compressed root file system to generate a burnable image and encapsulate it to generate a software update firmware package; mirroring the software update firmware package to a pre- or real-time compiled upgrade service framework, and executing the upgrade service framework to upgrade the target software through the software update firmware package. The present application adopts a combination of an unordered block mirror file system and a compressed root file system to perform software upgrades. The read-only nature of the compressed root file system ensures the security of system files. Encapsulating the unordered block mirror file system and the compressed root file system in one partition can save storage space management overhead, greatly shortening the development cycle and improving robustness.

In accordance with an embodiment, described herein is a system and method for providing cross-microservice query processing. The system provides an object service framework that supports the use of microservices that may be loosely-coupled but related in some way, for example in that they interoperate together or require access to each other's data in order to process queries. Each microservice can be developed, deployed and evolve independently, and interact with the other microservices through contracts or interfaces that are defined as public APIs and are then exposed via the framework. The object service framework can be used, for example to provide a cross-microservice layer that automatically transforms queries that join objects in different microservices into a single database query that is optimized for use with the database.

The invention relates to a privacy protection-oriented cryptography credential model watermark service method and system, and the core of the invention is to initiate a privacy protection-based model watermark, i.e., a service framework: a service provider does not need to access private data and model information of a model owner at all, and uses a proxy model pool and a public data set of the service provider; and generating a watermark pair with strong mobility and robustness through iterative optimization of a watermark trigger image generator and a plurality of task agent models. According to the method, the digital signature and the credible timestamp are combined for the first time, a verifiable watermark voucher is generated, and the generation time, the service source and the integrity of the watermark are ensured not to be tampered and non-repudiated, so that ownership disputes such as model overwriting attacks and the like are effectively coped with. According to the method, three bottlenecks of privacy disclosure, poor mobility and insufficient legal evidence force in the existing watermarking technology are fundamentally solved, and the method is particularly suitable for ownership protection of image processing models.

The application provides a blockchain smart contract service framework design method based on a hash multi-way tree and an application system, relates to the technical field of blockchain smart contracts, and forms a hash multi-way tree on the blockchain by connecting a plurality of node smart contracts to each other with each node smart contract as a tree node, realizes the construction of a blockchain smart contract framework based on the hash multi-way tree, realizes the scalability and the integrated multi-way tree structure on the blockchain, internally builds an access mechanism based on roles and permissions for each tree node of the hash multi-way tree, checks whether a user has specific roles and permissions, sets a user access interface externally, provides a serialization engine and a deserialization engine, realizes two functional domains of data management and contract management, and can effectively improve the scalability, compatibility, access speed, maintainability, data confidentiality and data integrity of a DApp.

The present invention discloses a secure edge server system for industrial early warning, which is composed of an interface layer, a data layer, an application layer, and a digital twin layer in the Baetyl / Link IoT Edge / ROS 2.0 architecture to build an edge collaborative service framework. By establishing a digital virtual model, the present invention not only realizes advanced fault warning reminders, handles hidden dangers in equipment in advance, and improves equipment utilization, but also has a built-in statistical system that records historical data and combines it with the pre-simulation operation data of the digital twin layer. According to data analysis, it derives the optimal control parameters for equipment operation, or judges various key production data during equipment operation based on data inflection points, thereby improving the equipment management level of the production line and the control level of the production process. At the same time, it conducts comprehensive and integrated situational awareness and monitoring and early warning from aspects such as personnel, equipment, environment, and network systems, and enhances the self-protection capability of the monitoring and early warning platform.

The invention relates to a text processing method and apparatus, a device, a medium and a program product. The method comprises the steps of obtaining an identifier of a target model and a target dialogue text carried by a model reasoning request; under the condition that the first type of dialogue template is built in the target model, converting the target dialogue text into an input sequence by utilizing the first type of dialogue template built in the target model; under the condition that the target model is not internally provided with the first type of dialogue template and the second type of dialogue template supports the target model, converting the target dialogue text into an input sequence by utilizing the second type of dialogue template; and under the condition that the target model is not internally provided with the first type of dialogue template and the second type of dialogue template does not support the target model, converting the target dialogue text into an input sequence by utilizing the first type of dialogue template predefined outside the target model. Therefore, the target model service platform can support the dialogue service of models adopting different dialogue templates under the target model service framework.

The invention relates to a service processing method based on thread pool management, and belongs to the technical field of Internet of Things. According to the method, firstly, public basic services of the power Internet of Things are integrated into a basic service framework, a unified message interface is set for the basic service framework for a user to develop services, when a user application is started, service initialization is carried out, and after initialization is finished, periodic operation is carried out, and service message processing is carried out; and when it is detected that the number of messages needing to run in the thread pool exceeds an expected value, early backlog messages in the thread pool are cleared. According to the invention, based on the basic service framework of the power Internet of Things, the thread pool is uniformly adopted to schedule and manage service processing, and when the number of messages in the thread pool is detected to exceed an expected value, earlier backlog messages are cleared, so that excessive backlog of the messages caused by massive messages and influence on normal message processing are prevented, the execution processing efficiency of the service is improved, and the user experience is improved. And the security, fault tolerance and stability of the application are improved.

The invention discloses a method for assisting anti-fraud algorithm iteration based on an AI agent, and the method comprises the steps: building a rear-end service framework through a network interface based on an MCP technology; constructing a system cue word for guiding the LLM to generate an output conforming to expectation; testing the capability of the LLM for performing complex task planning according to the natural language instruction until a set requirement is met; the back-end service framework analyzes the input fraud-related request, splices the input fraud-related request with a system cue word and then transmits the spliced fraud-related request and the system cue word to the LLM; calling of all tools in the back-end service framework is automatically planned through the LLM according to the received content; and generating a fraud identification result through the LLM according to an output result of each tool in the back-end service framework. According to the method, fraud identification is performed through an automatic incremental training process and a plurality of identification models, so that the identification capability of the identification models is enhanced, the fraud identification precision is greatly improved, and fraud information is effectively intercepted.