31 results about "Context recognition" patented technology

The invention discloses a bill recognition method and system based on artificial intelligence image enhancement, and relates to the field of image recognition. The method comprises the following steps: S1, extracting multi-dimensional quality features based on an original image of a bill and calculating a scene consistency factor; s2, adjusting a global enhancement weight according to a scene consistency factor, adjusting a local gain in combination with a detail fidelity factor, and performing adaptive enhancement on the original image to generate an enhanced image; s3, establishing an optical flow field model to analyze geometric deformation of the enhanced image, and performing adaptive correction in combination with local deformation rigidity to generate a corrected image; and S4, analyzing gradient features and character confidence of the corrected image, extracting a candidate character region, and performing context recognition by adopting a sequence model to obtain a text field set. Scene complexity is quantified through multi-dimensional quality features, and detail fidelity self-adaptive enhancement and optical flow deformation correction of high-frequency character distinguishing are combined, so that bill character definition and recognition accuracy are remarkably improved.

The invention provides a chat question-answering method and device based on knowledge precipitation, equipment and a storage medium, relates to the technical field of data processing, in particular to the fields of artificial intelligence, domain knowledge modeling and the like, and can be applied to application scenes such as chat conversation intelligent question-answering and the like. The specific implementation scheme is as follows: in response to a current message event in a chat dialogue, determining context information of the current dialogue; performing intention recognition based on the context information to obtain an intention recognition result; retrieving related domain knowledge from a knowledge base based on the intention recognition result; using the retrieved domain knowledge and context information to generate candidate answers; and carrying out quality scoring on the candidate answers, screening the answers with scores reaching a preset threshold value as target answers, and sending the target answers to the chat conversation. According to the scheme, the context can be constructed in the chat conversation, the intention can be recognized, the precipitated knowledge can be retrieved, and the high-quality answer can be generated, so that real-time reliable reply is provided, efficiency and accuracy are improved, and knowledge reuse and dynamic update are promoted.

An information processing device (1) comprises a reception unit (31), a determination unit (34), an identification unit (35), and a provisioning unit (36). The reception unit (31) receives a search query from a user belonging to an organization. The determination unit (34) determines a user context on the basis of user information and the search query received by the reception unit (31). The identification unit (35) identifies, as a target user, another user who belongs to the organization and who has a past or current context that is similar to or matches the user context determined by the determination unit (34). The provisioning unit (36) provides information pertaining to the target user identified by the identification unit (35).

The invention relates to the technical field of data transmission, and particularly provides a streaming Token double-path transmission method, system and device and a storage medium, and the method comprises the steps: responding to a streaming request, completing the initialization of a system and a session, building a bypass channel from a model to a client, and creating a session context in a control flow engine; identifying the type of each Token generated by the model, and shunting according to the type: data Token routing to a bypass channel and control Token routing to a control flow path processed by a control flow engine; data bypass transmission and control flow service processing are executed in parallel, and the states of the data bypass transmission and the control flow service processing are ensured to be consistent through a state synchronization mechanism; and the dual-path operation state is continuously monitored, and dynamic path adjustment or abnormity recovery is carried out when abnormity is detected. According to the method, the data flow and the control flow are decoupled, so that the transmission delay of the streaming Token is remarkably reduced while the state consistency is ensured.

The application discloses a tourist sentiment analysis method based on a large language model, and belongs to the technical field of intelligent scenic spot management, and solves the problems of poor real-time performance, low Chinese context recognition accuracy, insufficient multi-modal fusion, lack of multi-region sentiment linkage prediction and closed-loop management capability in the prior art. The core steps of the method include: S1 data acquisition and preprocessing, S2 basic sentiment calculation, S3 Chinese context enhancement correction, S4 final sentiment value fusion calculation, S5 regional sentiment index construction and early warning, S6 closed-loop feedback regulation, and the S7 multi-region sentiment coupling prediction and pre-intervention step is additionally added. The application supports text and voice multi-modal data acquisition, designs a special sentiment correction mechanism for Chinese context, can realize real-time analysis of tourist sentiment, negative emotion early warning and pre-intervention, and forms an operation optimization closed loop, and provides accurate decision basis for fine management of intelligent scenic spots.

Systems and methods for transferring link context to a web browser. An example method includes receiving, by a web browser, a web address from a desktop application based on a selection of a link in document displayed in the desktop application; receiving, by a web browser from the application, link context for the selected link, the link context including an identification of the desktop application and an identification of the document; loading a web page corresponding to the received web address; based on the link context, identifying a browser-based application corresponding to the desktop application; launching the browser-based application within the web browser; transmitting the link context to the browser-based application; and displaying at least a portion of the document within the web browser.

The application relates to the field of data processing, and provides a multi-process DMA access method and device based on device end context recognition and a chip. The method comprises the following steps: when initiating a DMA operation, a target VF generates a DMA request containing a local virtual address of the target VF; a VFID corresponding to the target VF is extracted from the DMA request, a virtual function address conversion table in a terminal device is inquired according to the VFID, and an address mapping table is acquired; the local virtual address of the target VF is converted into a corresponding DVA through the address mapping table; a standard PCIe TLP is generated based on the converted DVA; and the TLP is sent to a host, so that an input / output memory management unit or an address conversion service on the host side converts the DVA into a host physical address HPA, and the DMA operation is completed. Through the front movement of the multi-process address conversion from the host side to the terminal device side, a more efficient DMA access path is realized without relying on a PASID.

A system can comprise a memory that stores computer executable components, and a processor that executes the computer executable components stored in the memory. The computer executable components can comprise a correlation component that, based on a first context defining a first spare resource instance to be utilized in a networked system, identifies a second spare resource instance, and an evaluation component that generates an availability score defining prediction of availability of the first spare resource instance based on a second context defining the second spare resource instance, wherein the evaluation component further deploys the first spare resource instance, based on the availability score, in the networked system.

Disclosed are methods, systems, and computer-readable medium to perform operations including: receiving coded data comprising a plurality of data elements, each data element comprising a coded message; determining, for a coded message and based on a semantic model, a semantic distortion value, the semantic model comprising one or more classifiers configured to identify a desired meaning of a given coded message based on a context of the given coded message; determining a difference between the semantic distortion value and a baseline semantic distortion value; and generating, when the difference exceeds a threshold value, a security message indicating a potential security issue.

The invention relates to the technical field of voice recognition, in particular to an intelligent telephone traffic system which comprises a signal verification module, a voiceprint separation module, a content transfer module, an intention chain construction module and a priority judgment module. According to the method, in the voice input process, stability verification is achieved through comparison of the breathing flow velocity curve and the fundamental frequency period, it is guaranteed that recognition signals are consistent, breathing fluctuation deviation is reduced, acoustic features are corrected through alignment of the oral formant and the voiceprint template, and the voice is closer to real voice; multi-source scene analysis formant position and bandwidth change improves sound channel difference to avoid feature aliasing of speakers, phoneme boundaries are subjected to difference value and sequence rearrangement to realize stable time sequence matching to ensure consistent transcription, intention processing is combined with verb object time chain arrangement to generate distinguishable priorities in conflict judgment to ensure multi-intention semantic coherence and accuracy, and multi-intention semantic coherence is realized. Input stable signals are integrally formed to distinguish time sequence optimization intention sequencing continuous links, and the complex context recognition precision and logic rationality are improved.

This invention relates to a method and system for automatically generating prompts for agents in a PLM system. The method includes: upon receiving a prompt generation task from a user, constructing a structured field-level context based on key fields of the current page; identifying the current business scenario and loading an agent task flow matching the current business scenario; multiple agents are configured for different functions, and the agent task flow consists of the agents to be called and the execution order between them; generating instructions based on the field dimensions of the business scenario, agents, and field-level context, combined with context information, and inputting the instructions into the agent task flow to generate prompts; by constructing a unified field context acquisition and prompt template instantiation mechanism, driving multiple domain agents to work collaboratively according to business scenarios, significantly improving the business accuracy, technical depth, and overall efficiency of agents in scenarios such as engineering changes and quality analysis.

The system according to this embodiment aims to detect slang related to crime from posts on social media and identify accounts that are highly likely to be involved in criminal activity. [Solution] The system according to the embodiment comprises a slang detection unit, a context analysis unit, and an account identification unit. The slang detection unit detects slang related to crime from posts on social media. The context analysis unit analyzes the context before and after the slang detected by the slang detection unit. The account identification unit identifies accounts that are highly likely to be involved in criminal activity based on the context analyzed by the context analysis unit.

The present invention provides a method and apparatus for performing cognitive digital twinning using an advanced spatial context recognition and reasoning process based on a multi-dimensional facility information model. Based on multi-dimensional facility information for each of a plurality of facilities, a three-dimensional virtual space is implemented for each of the plurality of facilities, and a cell matrix is set for each of a plurality of planes, based on which a cell space list is set for each of a plurality of structure information, a plurality of facility resource information, and a plurality of vector space information, and the 3D virtual space for each of the plurality of facilities is a vector space and a cell represented in space, and based on information about a plurality of sensors pre-installed in each of the plurality of facilities and the multi-dimensional facility information.

The invention relates to the field of data processing, and provides a multi-process DMA (Direct Memory Access) method and device based on equipment-side context identification, and a chip. The method comprises the steps that when DMA operation is initiated, a target VF generates a DMA request containing a local virtual address of the target VF; a VFID corresponding to the target VF is extracted from the DMA request, a virtual function address translation table in the terminal equipment is inquired according to the VFID, and an address mapping table is obtained; converting a local virtual address of the target VF into a corresponding DVA through an address mapping table; generating a standard PCIe TLP based on the DVA obtained by conversion; and the TLP is sent to the host, so that an input / output memory management unit or an address conversion service at the host side converts the DVA into a host physical address HPA, and DMA operation is completed. Multi-process address translation is moved forward from a host side to a terminal device side, and a more efficient DMA access path is realized on the premise of not depending on PASID.

This invention provides a distributed storage message interaction method, system, device, and storage medium. The method reduces memory consumption in the IO stack and improves memory utilization efficiency through IO context integration; it converts message transmission to semi-asynchronous mode through context integration, reducing CPU consumption; it reduces file descriptor usage during message interaction through time-division multiplexing of sockets; it improves overall network bandwidth utilization through socket multi-channel technology; it reduces the impact of data retransmission on the system through IO context identification; and it accelerates IO processing and message processing speed through polling mode, thereby reducing the usage of CPU, memory, and file descriptors in the entire HCI environment, while improving IO speed and reducing IO latency in the entire HCI scenario.

The application discloses a document historical context extraction method and device based on information extraction, and relates to the field of information extraction.The method comprises the following steps: obtaining a document, and performing cleaning processing on the document according to a set character cleaning rule to obtain a cleaned document; analyzing the content of clauses of the cleaned document, and performing hierarchical representation on the document according to the hierarchical serial numbers of the clauses; identifying the clause relationship and the clause time of the document based on a deep learning model, and outputting an identification result.The application can greatly improve the generalization and accuracy of the model, and can obtain strong generalization of the document context recognition ability only by using a small amount of labeled document files.

Technologies and techniques for detecting and identifying objects within a vehicle interior are disclosed. One or more cameras capture image data of the vehicle interior, which is processed and analyzed using a vision-language model (VLM) to detect and identify objects based on their visual and contextual characteristics. The system associates the identified objects with locations inside the vehicle and generates notifications containing the object details and locations. The detection criteria are dynamically updated based on contextual factors, such as vehicle location, environmental conditions, and user preferences. The system further adjusts future object detection criteria based on feedback received from users, enabling improved detection accuracy. Additionally, the system can identify partially obscured objects using image segmentation and contextual recognition techniques. Notifications are communicated through an interface or connected mobile devices, allowing users to interact with the detected objects and receive real-time updates on their locations.

The present invention discloses a generative AI system for real-time services combining a high-performance lightweight language model based on H-MoE and RAG, executable in an on-device CPU environment, which can provide smooth and reliable real-time services in an on-device CPU environment by applying high-performance lightweight LLM-RAG technology and configuring the model as H-MoE, and providing answers or results through an appropriate Expert Model based on multimodal context recognition. The present invention comprises: an LLM module equipped with an internal database based on multimodal data including text, sensor data, and image data, and external knowledge data based on RAG, wherein semantic-based search is performed by embeddings; and a dynamic emotion recognition module in which the influence of elements constituting a space of lighting, color, and sound on user emotion is analyzed based on natural language understanding and sensor data of the LLM module.

A method (1500) includes: detecting (1502) a stroke input; identifying (1504) the shape and color of the stroke input; generating (1506) a plurality of first candidate objects based on the identified shape and color of the stroke input using an AI model; determining (1508) depth information of one or more objects to be inserted in the image; identifying (1510) a context of the image; identifying (1512) a location of the stroke input within the image, selecting (1514) one or more second candidate objects based on the identified context and the identified location, and updating (1516) the image by inserting the selected one or more second candidate objects based on the determined depth information of the stroke input and the identified location.

Systems and methods of the present disclosure are directed to a computing system. including one or more processors and a machine-learned multi-mode speech recognition model configured to operate in a streaming recognition mode or a contextual recognition mode. The computing system can perform operations including obtaining speech data and a ground truth label and processing the speech data using the contextual recognition mode to obtain contextual prediction data. The operations can include evaluating a difference between the contextual prediction data and the ground truth label and processing the speech data using the streaming recognition mode to obtain streaming prediction data. The operations can include evaluating a difference between the streaming prediction data and the ground truth label and the contextual and streaming prediction data. The operations can include adjusting parameters of the speech recognition model.

According to one embodiment, a method, computer system, and computer program product for selective image processing. The embodiment may include receiving a digital image for selective image processing. The embodiment may include identifying a context of the selective image processing. The embodiment may include performing object detection within the received digital image to identify one or more depicted objects based on the identified context. The embodiment may include overlaying, on at least one portion of the received digital image, one or more bounding boxes which encompass relative positions of the identified one or more depicted objects. The embodiment may include performing the selective image processing exclusively within the one or more overlaid bounding boxes.

This application relates to the fields of software engineering and system architecture design, and provides a method, system, and computer-readable storage medium for boundary context identification based on business semantic analysis. The method includes: performing semantic layer reasoning on an acquired business requirement document to obtain business semantic elements; mapping the business semantic elements of the business requirement document to nodes and connecting edges according to preset graph construction rules to obtain a business semantic coupling graph; identifying high-cohesion node sets in the business semantic coupling graph to form communities using graph algorithms, obtaining the initial boundary context corresponding to each community; applying heuristic rules to perform dynamic graph operations on the initial boundary context to obtain the final boundary context; and generating a visual interface to display the final boundary context and the set of interaction relationships between different final boundary contexts. This method automatically discovers boundary contexts in the early design stage based on business requirement documents, without relying on existing code and text similarity, ensuring the accuracy of automatic context identification.

Method and system for implementing RAG with improved context identification from semi-structured documents is disclosed. A target document is preprocessed by segmenting into individual pages. Text content from each individual page is extracted and processed through a domain-specific token filter to identify terms matching a stored domain-specific dictionary, and generate filtered page content comprising the matching terms. A page vector representation of the filtered page is computed, and vector similarity is performed between each page vector representation and stored vectors in a vector database that are associated with predefined entity types via entity labels that are extracted from matched templates. The entity labels are mapped to corresponding document pages based on similarity scores. A user query is analyzed using a Large Language Model (LLM) to identify relevant entity labels and pages with the relevant entity labels to generate a response.

The invention discloses an electroencephalogram adaptive filtering method and system based on multi-modal behavior perception, and the method comprises the steps: collecting multi-modal sensing data and electroencephalogram signals of users, extracting multi-modal features, and combining with a personalized feature vector which is allocated to each user in advance and reflects the weight credibility of different modal data; constructing a multi-modal fusion feature; inputting the context state information into a pre-trained context sensing model to obtain context state information; and extracting electroencephalogram signal features, fusing the electroencephalogram signal features with the context state information to obtain electroencephalogram fusion features, inputting the electroencephalogram fusion features into a pre-trained filtering parameter prediction model, and predicting filtering parameters. According to the method, context recognition pertinence and filtering parameter prediction accuracy are improved through personalized adaptation and double-fusion feature design, artifacts can be effectively removed, pure electroencephalogram signals are reserved, user adaptability, the denoising effect and application reliability are considered, and the method is suitable for scenes of brain-computer interfaces, neuroscience monitoring and the like.

The present application relates to a kind of enterprise-level intelligent text proofreading system, belong to enterprise data management technical field, the technical problem that the quality of the text of the existing method enterprise knowledge management library is difficult to be effectively guaranteed in warehouse is solved.Enterprise knowledge management library is input and retrieved module, for receiving text to be proofread and supporting multi-format input and draft temporary storage;Text context recognition module, utilize the text context recognition model of well-trained identification context category to which the text to be proofread belongs;Custom rule configuration module, for custom rule and dynamically loaded to intelligent text proofreading module;Intelligent text proofreading module, utilize the well-trained multi-pass mutual proofreading model, the context category and custom rule, proofread the text to be proofread, obtain proofreading result;Proofreading result output module, for the word-by-word weighted voting of proofreading result, generate final proofreading text.The high-quality content maintenance of the knowledge text of enterprise knowledge management library is realized.

Systems and methods are provided herein for determining types of references within a content item and mapping them to particular applications. A content management application identifies an entity and a context of the entity at a location within the content item. The content management application may identify the entity and the context of the entity in real time as a first user device processes the content item, or the content management application may identify and store the entity and the context of the entity in a database before providing the content item. After determining a presence of a second user device associated with a profile, the content management application determines at least one application associated with the entity and the context of the entity on the second user device and launches the application to create an immersive content consumption experience.

This invention belongs to the field of multilingual design optimization technology, and particularly relates to a method and system for automatic adaptation of multilingual design documents based on a cloud-based thesaurus. It constructs a multilingual cloud-based thesaurus to achieve versioned storage and real-time synchronization for multiple languages; uses a file parsing engine to perform semantic analysis and UI context recognition on the design documents to be adapted, constructing a semantic fingerprint containing multi-dimensional features; based on the multilingual cloud-based thesaurus, it employs a three-level matching algorithm combining forced terminology matching, semantic similarity matching, and layout adaptation matching to replace the design documents to be adapted with the target language version; and performs parallel visual preview of the replaced multilingual versions. This invention uses a three-level matching algorithm based on terminology priority, semantic similarity, and layout adaptation to achieve accurate multilingual replacement of design text, realizing for the first time the full-process automation of terminology management, accurate replacement, and visual evaluation.

An original email text is evaluated against a set of properties to identify shortcomings in the natural language, format, and structure used in the original email text. The properties represent characteristics needed to enhance the original mail text to a professionally-written style. A first machine learning model identifies the missing properties or characteristics in the original email that exist given a few-shot context. The few-shot context includes labeled email samples containing the desired properties. A second machine learning model is used to generate an enhanced email given the original email, personal data relevant to the content of the original data, the missing properties and the few-shot context.

The invention provides a command processing method and device, electronic equipment and a storage medium, and relates to the technical field of computers, in particular to the technical field of natural language processing and the like. According to the scheme, in response to a detected to-be-executed service command submitted to a command line environment, context recognition is carried out on the to-be-executed service command through a pre-injected command interception script, so that context information of the to-be-executed service command is obtained; based on effective configuration information and / or context information associated with the command interception script, performing enhanced arbitration on the to-be-executed service command to obtain an enhanced arbitration result of the to-be-executed service command; wherein the enhancement arbitration result is used for judging whether to execute enhancement processing on the to-be-executed service command; in response to the enhancement arbitration result indication, performing enhancement processing on the to-be-executed service command, and enhancing the to-be-executed service command to obtain an enhanced to-be-executed service command; and executing the enhanced to-be-executed service command.

The invention discloses an enterprise digital transformation index construction system based on a knowledge graph and an NLP, and relates to the technical field of enterprise digital transformation, and the enterprise digital transformation index construction system comprises a context source labeling module, a semantic segmentation extraction module, a semantic redirection module, a cross-domain semantic fusion module and a semantic weight balance module; and the context source labeling module is used for establishing a context source labeling system according to semantic differences between internal files and external reports of an enterprise, labeling a writing background and an application field of each text data record, and generating a context recognition basis in a text input stage. According to the method, semantic boundaries of different source texts are kept by introducing context source labels and combining semantic segmentation and semantic redirection, and the semantic understanding accuracy is improved. And meanwhile, through context consistency constraint and semantic weight dynamic balance, index fluctuation caused by semantic drift is inhibited, and the stability and credibility of the enterprise digital transformation index are enhanced.