133 results about "Abstraction layer" patented technology

The invention discloses a quantum classical hybrid computing task dynamic scheduling system and method based on SLURM. The system comprises a quantum task feature extraction module, a dynamic priority evaluation module, a dependency analysis module, a quantum perception backfilling module and a uniform resource abstraction layer module. The method comprises the following steps: extracting quantum features of a to-be-processed task and a computing resource to form a quantum feature set and caching the quantum feature set; the feature set is obtained in real time, task priorities are output through multi-dimensional evaluation, and a real-time priority sequence is generated; based on the task type and the feature set, forming a dependency relationship between the classic task and the quantum task, and converting the dependency relationship into a dependency constraint and / or resource reservation instruction; according to the task priority, the resource reservation instruction and the real-time resource state, a future idle period is predicted, and a short-time quantum task is inserted for backfilling; computing resources are distributed according to the task priority and the dependency constraint, and the resource utilization state is fed back to the backfill and priority evaluation module in real time. According to the method, efficient scheduling of hybrid computing tasks can be realized, and the overall performance of the system is remarkably improved.

A system and method for real-time or near real-time multilingual language transformation in telecommunications environments using distributed artificial intelligence (Al). The system includes at least one processor configured to instantiate a plurality of Al agents, each corresponding to a user in a communication session, and to manage their operation via an Al-agent runtime adapter comprising a microkernel agent manager and a virtual machine layer. A language model abstraction layer enables access to language models and Al algorithm libraries through publish-subscribe interfaces. The system supports contextual transformation of speech based on user profile data such as sentiment, emotional tone, and cultural background. Execution can occur on telecommunication-class switches, edge devices, or general-purpose computing hardware. Al agents operate asynchronously, access shared or private memory, and produce personalized, bidirectional linguistic output. The architecture supports secure, scalable deployment across heterogeneous devices and networks.

The present application provides a large language model-based query statement generation method, apparatus, and device, and a medium. The large language model comprises a plurality of encoders. The method comprises: acquiring a query text; performing vectorization processing on the query text to obtain a target vector representation corresponding to the query text; inputting the target vector representation into the plurality of encoders, to perform encoding processing on the target vector representation by means of weight matrices of the encoders, so as to obtain context vectors corresponding to the encoders, wherein different encoders have different weight matrices; and generating a target query statement on the basis of the context vectors corresponding to the encoders. A model is allowed to focus on different information features in different attention heads, independently capture different aspects of a query text, and process information at multiple abstraction levels, thereby better capturing complex and abstract semantic relationships, effectively distinguishing the importance of information when processing complex and fuzzy questions, improving the accuracy of query statement generation.

The invention relates to code verification, in particular to a code and document consistency reasoning verification system based on a cross-modal logic atlas, which comprises an input layer for receiving source codes and software documents; the feature extraction layer is used for performing feature extraction on the source code and the software document; the abstraction layer is used for converting the extracted features into comparable intermediate representations; the unified modeling layer is used for mapping code logic and document semantics into a unified atlas structure and performing formalized representation by using description logic to obtain a cross-modal logic atlas; the reasoning verification layer is used for performing reasoning verification on the consistency of the code and the document according to the cross-modal logic graph; the result processing layer is used for judging a verification result according to the reasoning verification condition, starting a contradiction positioning mechanism when inconsistency is found, and carrying out contradiction positioning and report generation to obtain an interpretable report; according to the technical scheme provided by the invention, the defects that the consistency between the code and the document is difficult to accurately verify, and accurate contradiction positioning and an interpretable report are difficult to provide can be effectively overcome.

The invention discloses an intelligent routing and unified adaptation method for a large language model, and the method comprises the steps: defining all access details of the model through a declarative configuration file, and achieving the zero-code access of the large language model without writing any code for a newly-added model; a completely consistent calling interface is provided for an upstream application, the isomerism of all downstream large language models is shielded, and a unified request and response abstraction layer is constructed; through a strategy engine and a JSON path technology, complex streaming response including content thinking is precisely processed, and intelligent analysis and content extraction are carried out; dynamic configuration and intelligent strategy hot update are supported; intelligent routing of the model is realized, and an optimal large language model instance is dynamically selected; and meanwhile, enterprise-level governance capability is provided, governance functions such as fusing, degradation, current limiting and monitoring are integrated, and stability guarantee is provided for model calling. Therefore, the maintainability, the expandability and the user experience consistency of the system are comprehensively improved.

A system and method of decoding a bitstream to produce tensors for use by a neural network second portion. The method comprises decoding a network abstraction layer (NAL) unit from the bitstream having a predetermined length, wherein the NAL unit of the predetermined length indicates a network abstraction layer (NAL) unit format of one inner codec of a plurality of inner codecs, each other inner codec having NAL unit lengths different to the predetermined length; selecting an inner codec from the plurality of inner codecs based on the decoded NAL unit of the predetermined length; and decoding the bitstream using the selected inner codec to produce the tensors.

The invention discloses a construction method for dynamically reconstructing a multi-modal edge computing hardware architecture and a cross-industry application method, and relates to the field of edge computing hardware architectures. A five-level heterogeneous architecture is adopted to construct a field self-adaptive dynamic reconstruction calculation unit, and an equipment abstraction layer is constructed to realize module hot plugging and calculation power distribution; a multi-protocol ubiquitous interconnection interface engine is established, a hardware-level analysis architecture is integrated, and multi-protocol analysis and rapid switching are supported; an environment self-adaptive energy efficiency regulation and control system is deployed, and wide-temperature operation is realized by adopting a bionic composite heat dissipation design; an edge-region-cloud three-layer elastic scheduling network is constructed, cross-domain computing power arrangement is realized through federal learning, and the method is adaptive to multi-industry scenes. According to the method, the hardware reuse rate and the wide-temperature environment adaptability are improved, the multi-protocol processing delay is reduced, the computing power scheduling efficiency is optimized, the limitation of traditional equipment is broken through, and the efficient and stable operation requirements of cross-industry scenes such as electric power and medical treatment are met.

The invention relates to the technical field of computer server resource management and network virtualization, and particularly discloses a server hybrid deployment and management system based on a virtualization technology, which comprises a uniform resource abstraction layer, a global resource sensing and modeling unit, a strategy-driven intelligent arrangement engine and a heterogeneous resource execution adapter. The uniform resource abstraction layer performs standardized abstraction on a physical server, a virtual machine and a container; the global resource sensing and modeling unit integrates the resource data and constructs a global resource model; a strategy-driven intelligent arrangement engine performs multi-objective optimization solution based on strategies and constraints to generate a scheduling scheme; the heterogeneous resource execution adapter translates the scheme into bottom executable atomic operations. According to the invention, unified management and intelligent cooperative scheduling of heterogeneous computing resources are realized, and the resource utilization efficiency and the system automation level are improved.

The invention discloses a data layout optimization method and device applied to NPU code compilation and a medium. The method comprises the steps that an intermediate representation IR input by an upper layer is split into an operation type OP and an operand, and the operand input by the upper layer is divided into logic data and a logic mask; converting the logic data of the upper layer into data vector representation of the abstraction layer, and converting the logic mask of the upper layer into mask vector representation of the abstraction layer; obtaining a corresponding bottom hardware instruction capability according to an operation type OP input by an upper layer, and performing legalization operation on the operation type OP; and performing legalization processing on the converted abstract data vector type representation and mask vector representation according to the target hardware capability, performing instruction mapping on the processed abstract instruction according to the underlying hardware pair operation type OP, and generating a target LLVM instruction compatible with the underlying hardware. Register layout abstraction and conversion can be carried out on the NPU code under mixed precision input, and the utilization efficiency of the NPU bottom layer register is improved.

Device for the intelligent encapsulation of legacy data information systems with the integration of artificial intelligence, comprising: a collection unit for legacy interfaces, configured to intercept communication data from existing platforms; a semantic interpretation unit designed to analyze and map data structures and process semantics; a contextual abstraction layer designed to convert interpreted information into standardized, interoperable formats; a policy-driven management module that ensures secure access, Data transformation and regulatory compliance are ensured; and a learning and inference unit for adaptive optimization of data mappings and operational logic.

The invention discloses a method and system for packaging a webpage embedded sign language translation JS library, and belongs to the field of electrical digital data processing.The method comprises the steps that by constructing an ECS core architecture with perception feedback, a bidirectional data flow between modules is established to achieve dynamic collaboration of rendering and grammar; predictive streaming resource management is adopted, and resources are intelligently preloaded based on gesture sequence contexts; multi-instance physical isolation is realized by utilizing a sandboxed container and a ShadowDOM (Document Object Model); implementing multi-level adaptive rendering, and dynamically switching a rendering path and a frame rate according to equipment performance; and finally, shielding the bottom layer difference through cross-end unified rendering of the abstraction layer. According to the method, the problems of animation and semantic dislocation, resource loading conflict and poor equipment compatibility in a multi-user concurrent scene are effectively solved, and high-fidelity, low-delay and cross-end consistent experience of sign language translation animation is realized.

The invention relates to the technical field of automatic data processing of integrated circuit design, in particular to an automatic parasitic parameter result analysis system based on an FUZZ test. The fuzzy test engine module adopts syntax tree variation to generate a test case library comprising illegal characters, extreme values and logic loops, the exception response module captures illegal operations in an analysis and simulation stage through the probe detection unit, and the correction decision unit executes data restoration or generates an exception report. And the incremental updating unit synchronizes the abnormal features to a model self-optimization module rule base. And the dynamic adaptation module converts the multi-process parasitic parameters into a standardized simulation model through the process parameter abstraction layer. All the modules form a closed-loop system for test case generation, anomaly detection and repair and rule dynamic optimization, automatic processing of an unexpected input scene is achieved, and robustness and multi-process compatibility of parasitic parameter analysis are improved.

The invention provides a server delivery and operation and maintenance automation system oriented to multi-platform collaboration, and aims to solve the problems of split delivery process, lagging expansion response, high operation and maintenance complexity and the like in a heterogeneous infrastructure environment. According to the system, by constructing a multi-platform adaptive abstraction layer, bottom layer differences of private cloud, job scheduling platforms and the like are shielded, and unified modeling and standardization of delivery and operation and maintenance processes are achieved; on the basis, key links such as installation file generation, expansion file generation, configuration injection and deployment verification are integrated, an end-to-end one-key delivery assembly line is built, and rapid and consistent deployment of a cross-platform environment is supported; meanwhile, an automatic operation and maintenance closed loop is constructed, and operation and maintenance operations such as health examination, fault warning, configuration repair, service restart, node replacement and elastic capacity expansion and contraction are automatically executed by collecting operation states, logs and performance indexes of the server and the service in real time and combining a preset strategy or an intelligent analysis model. By means of the method, efficient delivery and intelligent capacity expansion of server resources are achieved, and the autonomous operation and maintenance capacity, the high availability and the service continuity guarantee level of the system in a complex multi-platform mixed deployment scene are remarkably improved.

The invention discloses a domain controller automatic test scheduling method and system based on a hierarchical software architecture, and relates to the technical field of automatic test and industrial control, and the method comprises the steps: S1, constructing an automatic test system of the hierarchical software architecture which at least comprises a hardware drive abstraction layer and a test application layer, the hardware drive abstraction layer is used for uniformly packaging bottom layer drives of a board card, a signal conditioning module, a relay matrix and a power supply module in the test system, and providing a standardized interface irrelevant to specific hardware for the test application layer; according to the domain controller automatic test scheduling method and system based on the hierarchical software architecture, the comprehensive requirements of a new energy automobile power domain controller on high-efficiency, high-reliability and high-consistency automatic test in research, development, verification and mass production stages are better met.

The invention discloses an RoCEv2 protocol RC service type basic primitive function verification system based on a UVM. The system comprises a test case class used for configuring verification parameters, initializing a verification environment and driving a primitive execution process; the verification environment is used for packaging a plurality of agent components, a reference model, a register model and a scoreboard and carrying out data interaction through a TLM interface; the agent components are used for driving or monitoring an interface designed to be tested, comprise but are not limited to a register read-write agent, a work queue element WQE reading agent, a load data reading and receiving agent, a data packet sending and receiving agent, a complete queue element CQE receiving agent and a register model component, and are constructed based on a register abstraction layer RAL modeling method; and front-door and back-door access paths are supported. The method is high in automation degree, integrates constraint randomization, comparison verification and coverage rate analysis, and realizes full automation of the verification process.

The invention relates to the technical field of digital twin and multi-source heterogeneous data fusion, and discloses a method for loading and rendering massive homologous heterogeneous data of a digital twin platform. The method aims at solving the problems of low loading efficiency, large rendering fluctuation, data semantic segmentation and the like in the prior art. The technical scheme comprises the steps of preprocessing data and establishing a correlation index; during initialization, a network state is adapted to load basic data; constructing unified abstraction layer standardized layer management; on-demand loading and multi-level caching are realized through the intelligent scheduler; a fusion rendering pipeline is configured to guarantee stable output of the 60FPS; operating a service binding engine to realize bidirectional linkage; measurement, model import, file preview, layer linkage and data maintenance are supported. The mass data loading efficiency is improved, memory overflow is avoided, service collaboration and system stability are enhanced, and the method is suitable for scenes such as urban planning, industrial simulation and emergency command.

The present invention provides a device, a computer-readable storage medium, and a method for exchanging data between containers, all equipped with shared memory. [Solution] Embodiments of this specification describe a Platform Abstraction Layer (PAL) for edge devices that enables services developed in different programming languages ​​to run as a pipeline on edge devices (e.g., cameras). For example, services may be deployed in separate containers developed using different programming languages. In one embodiment, stages in the pipeline can use buffers in shared memory to share data. The data may include a unique ID (e.g., frame ID) to identify the buffer containing the processed data by searching an index array for the shared memory. In this way, the shared memory supports random access.

The invention discloses a software architecture design method of an intelligence and calculation all-in-one machine, and relates to the field of artificial intelligence, the intelligence and calculation all-in-one machine comprises an AI application layer, an AI service layer, an AI platform development layer, an infrastructure abstraction layer and a security operation and maintenance layer, the AI application layer receives a task request of a user side, generates a task instruction and sends the task instruction to the AI platform development layer; the AI platform development layer receives a task instruction, obtains to-be-processed data and feeds back the to-be-processed data to the AI application layer, the AI application layer sends the to-be-processed data and a model demand to the AI service layer, the AI service layer loads a model, the loaded model processes the to-be-processed data and obtains a task processing result, the AI service layer sends the task processing result to the AI application layer, and the AI application layer sends the task processing result to the AI application layer. The AI application layer feeds back to a user side, the safety operation and maintenance layer monitors the operation process of each layer, and the infrastructure abstraction layer provides operation resources. The system is vertically divided into a plurality of different layers, the layers are in communication connection, the coupling degree is reduced, and the expandability is improved.

The application relates to the technical field of industrial measurement and control software, and discloses a nationwide state-oriented measurement and control platform which adopts a layered architecture and comprises the following: an application service layer for providing a man-machine interactive interface and executing business logic; a unified business message layer for providing a standardized data communication interface; a multithreaded time sequence scheduling engine for parallel receiving, scheduling and processing real-time tasks; a unified hardware abstraction layer for encapsulating the differences of different field bus drivers and providing a unified device operation interface; a domestication adaptation layer constructed on a Qt core framework and used for providing a unified driving abstraction across domestic chips and operating systems; and a hardware access layer for directly interacting with physical hardware devices.The application realizes 'one-time development and cross-domestic platform deployment', significantly improves the real-time performance, maintainability and expandability of the system under the domestic ecological environment, and meets the dual requirements of self-controllability and high performance.

The invention discloses a Direct3D stream output virtual implementation method based on kernel enhancement, which comprises the following steps: in a DXVK compiling stage, constructing stream data for a shader, establishing a format mapping table and an application scene library of D3D, DXVK and a kernel, initializing a shared memory pool when the kernel is started, after the DXVK is started, establishing an abstraction layer consistent with VKEXTTransformFEDback, analyzing parameters when an application calls a stream output related API (Application Program Interface), and implementing the kernel enhancement based Direct3D stream output virtual implementation. The buffer binding type API triggers a kernel to distribute partitions for streams, the stream control type API triggers a corresponding shader to rewrite, the query operation type API extracts query related information, the execution shader writes data into a temporary delivery buffer, data distribution is completed when a distribution condition is met, a distribution channel of an abstraction layer is packaged into a Vulkan secondary command buffer, and the Vulkan secondary command buffer is used for sending the data to the execution shader. And reusing a DXVK rendering thread command pool and submitting the command pool to the GPU in batches to finish rendering.

This invention relates to a multi-agent orchestration and execution method supporting complex cyclic graph structures, comprising: S1: receiving workflow configuration and constructing a directed graph data structure; S2: performing a depth-first search traversal on the directed graph data structure, detecting and outputting strongly connected components in the directed graph data structure; S3: abstracting strongly connected components containing multiple nodes into super nodes, and keeping single-node strongly connected components as ordinary nodes, thereby constructing a directed acyclic graph containing super nodes; S4: performing topological sorting on the directed acyclic graph to generate a hierarchical execution plan containing multiple execution layers; S5: scheduling execution items according to the hierarchical execution plan, and when an execution item is the super node, starting a loop executor for iterative execution. This method, by constructing a multi-layered technical system, integrates graph algorithm innovation, unified context management, abstraction layer design, metadata-driven techniques, etc., to achieve automated orchestration and execution of multi-agent workflows supporting cyclic graph structures.

This invention relates to an industrial shop floor machine learning system, comprising: a machine learning model providing machine learning data; an industrial shop floor providing shop floor data; and an abstraction layer connecting the machine learning model and the industrial shop floor, wherein the abstraction layer is configured to provide standardized communication between the machine learning model and the industrial shop floor using a machine learning markup language.

The application provides a verification method and related device for a graphics processor, comprising: abstracting, by an interface abstraction layer, original interface access for a target graphics processor into a function call for a virtual firmware execution module, wherein the virtual firmware execution module is a function level modeling of a control processor subsystem in the target graphics processor; the virtual firmware execution module generates a standard transaction sequence according to the function call, so that the virtual firmware execution module and / or the target image processor executes the standard transaction sequence; a consistency monitoring module obtains actual data corresponding to a plurality of checkpoints respectively, wherein the plurality of checkpoints correspond to a whole flow of the interface abstraction layer, the virtual firmware execution module and the target image processor; the consistency monitoring module compares the actual data corresponding to the plurality of checkpoints respectively with expected data corresponding to the plurality of checkpoints respectively one by one, and generates a verification report for the target graphics processor. In the verification process, no firmware instruction is relied on.

Technologies for creating an optimized and accelerated network pipeline using a shared interface Network Pipeline Abstraction Layer (NPAL) are described. A Data Processing Unit (DPU) comprises a physical port configured to couple with a breakout cable, which is physically coupled to a set of multiple devices, DPU hardware, and memory operationally coupled to the DPU hardware. The NPAL supports multiple logically split ports, with each logically split port corresponding to one of the multiple devices. The network pipeline comprises a set of tables and logic organized in a specific order to be accelerated by the acceleration hardware engine. The acceleration hardware engine is designed to process network traffic data using the network pipeline.

The invention relates to the technical field of virtual machines, and particularly discloses an equipment complete virtualization framework for a real-time operating system virtual machine (ZVM), which comprises a hardware platform, the ZVM and at least one client operating system, the ZVM is used for carrying out scheduling management and resource isolation on at least one guest operating system and providing at least one virtual device for each guest operating system; wherein the virtual equipment comprises a virtual equipment abstraction layer which is used for providing an equipment access interface for a client operating system; the virtual equipment simulation layer is used for intercepting and analyzing an equipment operation request initiated by a client operation system and simulating behavior logic of hardware equipment; and the data forwarding layer is used for forwarding part of the requests to hardware equipment of the hardware platform. By adopting the technical scheme provided by the invention, efficient multiplexing, safe access and real-time scheduling support of the same hardware equipment among multiple systems can be realized.

The invention discloses a cross-platform model deployment framework, electronic equipment and a storage medium, and the framework is applied to the electronic equipment and specifically comprises a framework interface abstraction layer and a plurality of runtime libraries connected with the framework interface abstraction layer. The framework interface abstraction layer is used for being connected with an application framework layer and providing a uniform abstraction interface for the application framework layer; the runtime library is used for providing running support for the framework interface abstraction layer. Compared with a traditional model deployment scheme, the scheme provided by the invention can accelerate the deployment speed of the model, ensures the uniformity of evaluation standards of the model on different platforms, and improves the reasoning performance. If the running environment is inferred to be updated, deployment updating of the model on a new platform can be rapidly completed according to the framework, and manpower and material resources needed by deployment can be greatly saved.

A system and method for dynamically selecting, executing, and optimizing microservices across multiple cloud service providers is disclosed. A workload management system receives a microservice execution request, retrieves metadata from a microservices catalog, and selects an optimal cloud provider using a weighted cost function based on execution performance, cost, security compliance, and availability. A machine learning model refines provider selection using historical execution data. An API abstraction layer formats and encrypts execution requests before transmission. A performance monitoring module tracks execution, while a fault detection module identifies failures and triggers automated failover via a failover module that preemptively replicates execution state data. A response processing module validates execution results before returning them to the requesting application. The system ensures dynamic, cost-efficient, and secure workload distribution while maintaining seamless failover resilience. By leveraging real-time performance metrics, historical data, and predictive analytics, the system optimizes microservice execution across diverse cloud environments.

The invention relates to a communication stack for energy storage communication and a data communication method of an energy storage system, and a layered software architecture comprises a plurality of function levels divided based on a microcontroller abstraction layer, an electronic control unit abstraction layer and a service layer in Autosar. The data communication module is used for adapting to the data communication requirement of an energy storage battery management system and opposite-end equipment for performing a target interconnection communication protocol; the standardized interface module is provided with a standardized interface framework for interaction among functional levels in the layered software architecture and interaction between the energy storage battery management system and opposite-end equipment, and supports setting of interface parameters based on data communication requirements so as to realize data interaction among the functional levels through the standardized interface framework and the set interface parameters. According to the embodiment of the invention, the design target of each function level is consistent with the positioning of the corresponding level of Autosar, so that the architecture logic of the communication stack is uniformly constrained, the software reuse rate and maintenance efficiency are improved, and the coupling degree of a bottom software module is reduced.

This invention discloses a method for intelligent routing and unified adaptation of large language models. It defines all access details of the model through a declarative configuration file, allowing new models to be added without writing any code, achieving zero-code access for large language models. It provides a completely consistent calling interface for upstream applications, shielding the heterogeneity of all downstream large language models and constructing a unified request and response abstraction layer. Through a strategy engine and JSON path technology, it accurately handles complex streaming responses, including thought content, with intelligent parsing and content extraction. It supports dynamic configuration and intelligent strategy hot updates. It achieves intelligent model routing, dynamically selecting the optimal large language model instance. Simultaneously, it provides enterprise-level governance capabilities, integrating circuit breaking, degradation, rate limiting, and monitoring functions to ensure stability for model calls. This comprehensively improves system maintainability, scalability, and user experience consistency.

The invention relates to the technical field of software development. The invention discloses a modularized code-free platform, which adopts a five-layer architecture and comprises an application presentation layer, a test report presentation layer and a test report presentation layer, the configuration engine layer is used for analyzing the structured configuration file and generating an executable process model; the test execution layer is used for scheduling and executing a test flow defined by the process model; the hardware abstraction layer is used for providing a unified instrument operation interface and managing a mapping relation between a logic signal and physical hardware; and the equipment driving layer is used for realizing communication with specific instrument equipment. According to the method, a configuration-as-a-service mode is adopted, so that business personnel can quickly construct an application on the premise of not writing codes, the technical threshold is remarkably reduced, and the response speed is increased; business personnel only need to edit Excel to complete configuration of a complex instrument control process, and codes do not need to be written.