206 results about "Language understanding" patented technology

A real-time system for the multimodal generation of virtual reality scenes based on artificial intelligence for the creation of immersive three-dimensional environments from natural language narratives, consisting of: a speech capture module configured to continuously record a user's spoken narrative via one or more directional microphones, preprocesses the captured signal by noise reduction and temporal alignment, and outputs a digital speech stream; A speech-to-text processing unit that is operationally coupled to the speech capture module and configured for real-time speech recognition using a continuous neural transformer model. The unit is trained to transcribe natural language utterances into structured text data while maintaining contextual continuity throughout the evolving narrative. a semantic interpretation processing unit that is communicatively linked to the speech recognition unit and configured to perform natural language understanding techniques to extract contextual entities, spatial references, temporal relationships, and object attributes from the transcribed narrative; the engine includes a large language model that is fine-tuned for spatial reasoning tasks; a scene graph generation module configured to transform the interpreted semantic data into a structured, hierarchical representation that defines nodes for identified entities and edges for corresponding relationships, with each node associated with metadata describing geometry, position, orientation, texture, and linking attributes between objects; a multimodal image-language model processor coupled with the scene graph generation module, wherein the processor is configured to retrieve, adapt, or synthesize appropriate three-dimensional elements from a pre-trained visual-lexical embedding space and align these elements with their semantic and spatial definitions derived from the scene graph; a scene assembly and rendering controller configured to create a cohesive virtual scene from the aligned assets, perform real-time rendering using a GPU-accelerated ray tracing pipeline, and produce a stereoscopic visual output that corresponds to the evolving narrative; A head-mounted virtual reality visualization device connected to the rendering engine and configured to display the generated immersive environment to the user in real time. The device features motion sensors and inside-out tracking cameras to detect head and body movements, dynamically updating viewing angles and perspective within the rendered scene; and a bidirectional feedback module integrated into the head-mounted device and connected to the semantic interpretation processing unit; the module is configured to interpret corrective commands, gestures, or supplementary comments from the user to refine or modify specific scene elements without interrupting the real-time visualization; The system continuously updates the virtual scene as the narrative develops, ensuring temporal synchronization between speech input and rendered output below a defined latency threshold, thus enabling a natural, dialogic construction of complex three-dimensional virtual environments.

The invention discloses a double-engine government affair question and answer method based on large model fine tuning and RAG retrieval, belongs to the field of government affair digitization and natural language processing, and combines large model language understanding generation ability, retrieval enhancement generation technology and a structured reasoning mode. The defects of a traditional government affair question and answer method in the aspects of dynamic policy response, complex semantic understanding and compliance control are overcome. Government affair field knowledge is adapted through large-model fine adjustment, and high-precision and timeliness answering of government affair consultation is realized in combination with vector retrieval and a dynamic updating mechanism. The core innovation of the method lies in deep fusion of a double-engine architecture and dynamic knowledge management, the accuracy and response efficiency of government affair questions and answers are improved on the premise of ensuring policy compliance, and the method is suitable for intelligent upgrading of scenes such as government affair service halls and online consultation platforms.

The invention relates to the technical field of artificial intelligence and Internet of Things, and discloses an AI interaction intelligent module based on a hybrid architecture, comprising a user interaction unit which supports voice, text and image multi-modal input and integrates intention recognition and context understanding algorithms; the data processing unit is used for carrying out structured processing on the electric appliance specification and the historical fault data and constructing a dynamically updated knowledge graph; the hybrid architecture core unit comprises a deep learning subunit for realizing natural language understanding and generation based on a Transform model, and a knowledge reasoning subunit; a fault diagnosis unit; and a feedback optimization unit. According to the method, seamless cooperation of deep learning and symbol logic is realized through a dynamic routing strategy, a high-confidence-coefficient scene generates a response through a Transform model, a medium-confidence-coefficient scene calls a knowledge graph rule for verification, and a low-confidence-coefficient scene supplements information through multiple rounds of interaction, so that the effect of improving balance efficiency and safety is achieved.

The invention discloses a multi-modal natural language understanding and generating system and method. The method comprises the following steps: constructing a cross-modal pre-training module, training a multi-modal encoder, and establishing a cross-modal association mapping space; mixing prompt fine tuning is carried out, and a complete blank filling template is constructed; according to the intention reasoning network, extracting multi-round dialogue intention representation of the user, and retrieving an external knowledge base for fine-grained reasoning; constructing a unified semantic representation framework, embedding the text, the image and the voice into a unified space, and generating a query vector of multi-modal intention perception; and the knowledge query module based on key value memory generates entity-level multi-modal replies and optimizes the semantic comprehension and generation capability of the dialogue model. According to the method, the multi-modal information understanding and generating capacity is improved, deep association and understanding of image and text information are achieved, downstream task adaptability is enhanced, task completion accuracy and efficiency are improved, unified semantic representation of the multi-modal information is achieved, and support is provided for information retrieval and utilization.

A computer-implemented method for operating an artificial intelligence voice agent system includes receiving voice input through communication channels; analyzing converted text through natural language processing (NLP) pipelines implementing intent recognition and sentiment analysis detecting emotional cues using a multimodal large language model (LLM); generating response content using machine learning models trained on domain-specific corpora; converting generated responses to synthetic speech through text-to-speech (TTS) engines; integrating with a customer relationship management (CRM) platforms or an enterprise resource planning (ERP) database; and implementing continuous learning by updating language understanding models using conversation logs, voice recognition parameters based on user feedback, and response generation patterns. One implementation is a computer-implemented system and method that operates a suite of intelligent interactive devices and platforms including an artificial intelligence voice agent, enhanced communication platforms, an intimacy companion system, and augmented / virtual reality eyeglasses. Further, one implementation includes AR / VR eyeglasses that project visual content onto interchangeable lenses or directly onto the user's retina via laser-based retinal projection, provide prescription adjustments, incorporate ear-mounted sensors for monitoring physiological parameters like heart rate, oxygen saturation, and blood pressure, and utilize wireless data transmission, onboard environmental sensing, and remote calibration, all designed to offer dynamically adaptive, secure, and context-aware interactions across communication, personal assistance, health monitoring, and immersive augmented or virtual reality environments.

Aiming at the problems of insufficient language understanding depth, weak personalized dialogue generation ability, lack of continuous learning and long-term user state modeling and the like in the current emotion recognition and psychological intervention technology, the invention provides a user emotion recognition and psychological intervention method combined with a large language model (LLM). According to the method, the potential emotional state is identified by analyzing free text information input by a user by utilizing the powerful capabilities of a large language model in the aspects of natural language understanding, emotional modeling and text generation; constructing a multi-round dialogue context, and reasoning a psychological change trend of the user; in combination with a psychological knowledge base, personalized and mild psychological intervention dialogue content with a dredging effect is generated. The system supports recognition and classification of various emotional states such as depression, anxiety and alonity, is suitable for various interaction scenes (such as APPs, webpages and social robots), and can greatly improve the precision of emotion recognition and the timeliness and effectiveness of psychological intervention. The emotion recognition and psychological intervention method based on the large language model provides solid technical support for constructing an intelligent, continuous and personalized psychological health management system, and has wide application prospects and profound social significance.

The invention relates to the technical field of electric digital data processing, in particular to an artificial intelligence algorithm-driven digital human language understanding and corresponding reaction interaction method. The method comprises the following steps: collecting historical language data, text data, action instruction data and physical environment information of a user to obtain structured training data; vectorizing the text data to obtain a text semantic vector; environment feature vectors are extracted from the physical environment information; splicing the text semantic vector and the environment feature vector to obtain an environment enhanced text vector; and constructing a pre-training language model based on the environment enhanced text vector, and performing secondary pre-training to generate an optimized language model parameter. According to the invention, by fusing the user language and the physical environment information and combining with the multi-module intelligent component, context perception, efficient understanding and multi-modal natural interaction of the digital human in a complex scene are realized, and the intelligence, adaptability and user experience of the system are remarkably improved.

In one embodiment, a method includes receiving training utterances associated with a domain, receiving ontology labels for the domain, wherein the ontology labels comprise one or more of an intent or a slot, generating an inventory for the domain, wherein the inventory comprises at least a respective index and respective span for each intent or slot, wherein the respective span comprises a respective descriptive label associated with the intent or slot, and wherein the respective descriptive label comprises a natural-language description of the intent or slot, generating frames for training utterances based on the training utterances and the inventory by a natural-language understanding (NLU) model, wherein each frame comprises a structural representation of the respective training utterance, wherein the structural representation is generated based on a comparison between the corresponding training utterance and the inventory, and updating the NLU model based on the frames.

The invention discloses an automatic data modeling and optimizing system and method fusing a knowledge graph and ChatBI, and relates to the technical field of data analysis. In order to solve the problems of high interaction threshold and insufficient analysis depth of a traditional BI tool, the scheme adopted by the invention comprises a data layer which has the capabilities of multi-source data access, domain knowledge graph construction, intelligent mapping recommendation, federal calculation and data mild governance, and realizes data integration and semantic unification; the analysis layer realizes accurate conversion from a natural language to an SQL and semantic reasoning of a knowledge graph through graph vectorization, multi-model cooperation, natural language understanding, intelligent SQL generation and graph dynamic updating, and supports efficient data query and analysis; and the application layer has the functions of natural language interaction, visual recommendation and generation, root cause analysis and intelligent early warning, and provides a visual interaction interface and data display service for a user. According to the invention, intelligent data analysis and visualization based on natural language interaction can be realized.

The invention provides a power industry-oriented language understanding and dialogue generation method, which comprises the following steps of: retrieving a dialogue process template related to a current problem from a pre-constructed power industry multi-round dialogue knowledge base according to a field label of the problem, and determining a basic framework of a dialogue process and key information required to be acquired by each round of dialogue; after each round of dialogue is finished, the obtained information is synthesized, reasoning is carried out through a pre-constructed electric power knowledge graph, the incidence relation between different professional fields is analyzed, key points and possible solutions of questions are deduced, and candidate answers are generated; and establishing an answer sorting model by applying deep learning, scoring and sorting the candidate answers, and selecting the answer with the highest score as the reply of the current round in combination with the correlation, integrity and specialty of the answers.

The invention relates to a pre-training language model fine tuning method, an inference service system, equipment and a medium, and the method comprises the steps: obtaining an original instruction of a target field and related fields, the original instruction comprising a language understanding task; evaluating and screening the original instruction according to a multi-dimensional index, and constructing a first high-value instruction set; extracting instructions in related fields from the first high-value instruction set, and performing filtering and mobility evaluation on the extracted instructions by using a pre-training language model to obtain a second high-value instruction set; adopting a pre-training language model to perform labeling, screening and enhancement processing on the second high-value instruction set in sequence, and generating a structured training sample; and training a pre-training language model according to the structured training sample to enable the pre-training language model to adapt to the target domain. Through the method and the device, the field annotation data quantity is reduced.

The invention discloses a document auditing method and device based on cooperation of a large model and a rule engine. The method comprises the following steps: analyzing an original document to form an intermediate representation file; extracting entities from the intermediate representation file to form an entity candidate set; inputting the thinking chain cue word into the large model to obtain a first triple, a first confidence coefficient and a first risk level, and calculating a first traceable score of the large model according to the reasoning path node; performing deterministic judgment by using the rule engine to obtain a second triple, a second confidence coefficient, a risk level and a second traceable score; calculating two weighted voting values and obtaining a conflict difference value; taking a conclusion corresponding to the high-weighted voting value as a final conclusion when the conflict difference value is smaller than a preset judgment threshold value; otherwise, starting an artificial rechecking process; and generating audit reports in various formats. According to the method, the advantages of relatively high language understanding ability of a large model and relatively high certainty of a rule engine are exerted, and the problem of missed checking or excessive marking is avoided.

The invention provides an intelligent customer service system, which comprises an information acquisition module used for identifying a customer identity according to a user login account and obtaining historical data under the account; the grading module is used for establishing a customer portrait model according to the historical data and classifying customer types; the matching module is used for automatically matching a service strategy model version corresponding to the account according to the customer type obtained by the grading module; the knowledge graph response module is used for carrying out deep analysis on client input contents based on the natural language understanding capability of a large language model in combination with a context, a client portrait and historical data after the version of the service strategy model is determined, automatically generating a task list and a semantic tag tree for a dialogue, and continuously optimizing the task list and the semantic tag tree; starting from a problem node of the task list, analyzing a corresponding semantic tag along an optimized and trained path in the knowledge graph, and matching a service processing action; and triggering a back-end system interface associated with the node according to needs, transferring real data, and solving client problems.

The invention discloses a data security classification and grading method and system based on a large language model, and relates to the technical field of language analysis, and the method comprises the steps: converting bank internal business data of each layer into text data through text conversion, unifying the data form of multi-layer or multi-modal data into text data, and storing the text data in a database; and a reliable basis is provided for subsequent language semantic recognition. According to the method, the text data is partitioned and sorted, subsequent context language understanding is facilitated, the context window length under each service scene is set, the context semantic understanding requirement of each service scene is met, and the reliability and accuracy of semantic association are improved. The internal business data of the bank is classified and graded based on the risk value of the business scene and the sensitive value of the composite field, the accuracy and adaptability of data security classification and grading are improved, security protection strategies of different degrees are carried out for different levels of different types of data, and the requirements of bank data storage and prevention are met.

The invention discloses a medical training evaluation method and system based on an intelligent simulation patient, and relates to the field of artificial intelligence technology and medical simulation, and the method comprises a virtual patient model construction step, a multi-modal perception step, an intelligent interaction step and an intelligent evaluation step. By integrating voice recognition, natural language understanding, facial expression recognition, gesture recognition and virtual reality technologies, a high-simulation and high-interactivity intelligent patient model is constructed for medical students or clinical medical staff to perform diagnosis and treatment training and skill operation assessment. The virtual patient model supports custom information such as age, gender, race, medical history, living habits and the like, different virtual patient images and different pathophysiological states are generated, and personalized learning requirements of users are met. In addition, targeted training suggestions are provided, the training difficulty and content can be adjusted according to the performance of the user, and different training requirements are met.

System and method for using knowledge fabric based on knowledge ontology, designed for deep language understanding and mechanistic causal reasoning, and meta-knowledge repository for auditable question answering. The method includes receiving an input text from a user, building a knowledge graph that represents real world facts and associations in the form of contextually tagged and weighted knowledge propositions, in multiple knowledge domains. The knowledge graph in combination with causal path knowledge and metadata describing digital sources containing answers constitutes the knowledge fabric. The method includes resolving ambiguity and determining actual intent of the user for the input text, from a plurality of interpretations of intent for sentences using the knowledge graph in conjunction with logical inference to achieve deep natural language understanding. The method includes finding / delivering response to the input request as to why / how unknown factors resulted in known outcome, or what outcomes are likely given known causal factors.

The invention discloses an accompanying dialogue system and method based on real-time environment perception and knowledge graph enhancement. The system comprises a multi-mode real-time environment perception module, a knowledge graph construction and enhancement module, a natural language understanding and dialogue management module, a personalized recommendation and narration generation module and an immersive multi-mode interaction module. The method comprises the steps of multi-modal real-time environment perception and situation data generation; performing dynamic association, query and enhanced reasoning on the knowledge graph; natural language understanding and dialogue management, personalized recommendation and narrative generation, and immersive multi-modal interaction presentation and feedback reception. According to the method and the system, the concern point of the user, namely scenery or details, can be accurately positioned, and context information required by subsequent service intelligence is provided, so that the problems of poor environment perception ability, weak interaction immersion, dull knowledge service, lack of individuation, insufficient intelligent accompanying experience and the like in the existing tourism auxiliary technology are solved.

The invention belongs to the technical field of contract review, and particularly relates to an intelligent contract review analysis method and system based on a large language model. According to the method, dynamic generation of rules and a neural symbol reasoning mechanism are mainly fused, and the core is to realize automatic and precise risk identification and evaluation of contracts by utilizing the powerful language understanding and generation capability of a large language model (LLM); the method comprises the specific steps of generating a contract review rule, extracting facts from a contract to be reviewed, performing review reasoning and generating a review report. According to the method, the inherent logic preciseness of symbol deduction and the powerful semantic understanding capability of a large language model are organically combined, so that core elements and potential risks in a contract can be more accurately identified, and contract terms which are complex in expression or have hidden agreement can be effectively processed and expressed.

The invention discloses a structured information extraction system for a mental disorder electronic medical record based on a knowledge graph and a natural language, and belongs to the technical field of medical information processing. According to the system, firstly, core concepts such as diseases, symptoms and drugs are extracted from an authoritative guide to construct a mental disorder knowledge graph, and a standardized clinical knowledge base is established; then pre-training and fine-tuning the deep learning model on a large number of biomedical texts and desensitized medical records to enable the deep learning model to have a medical language understanding ability; performing preprocessing, named entity recognition and entity linking on the electronic medical record text, and mapping spoken expressions to standard medical terms; inference is carried out by utilizing a relation extraction model and combining with a knowledge graph to complement implicit clinical information; and finally, structured data output meeting the standards of FHIR and the like is generated. According to the method, through deep fusion of knowledge driving and data driving, the problems of insufficient semantic understanding and weak generalization ability of a traditional method are effectively solved, and the accuracy and clinical value of electronic medical record structured information extraction are remarkably improved.

A public opinion analysis visualization platform construction method based on sentiment analysis belongs to the technical field of natural language processing, and specifically comprises the following steps: 1, collecting related corpora and data required by sentiment analysis; 2, preprocessing comment data required by sentiment analysis; and filtering invalid texts, and performing sentence segmentation processing. 3, firstly performing text classification on the processed data, then performing sentiment analysis, judging sentiment tendency in a comment text, and finally performing visual display; and 4, performing public opinion early warning based on an obtained sentiment analysis result and a popularity data analysis result. And 5, generating an independent public opinion evaluation report and a summary report through a related API interface of the local large language model. Through the method, the public opinion response efficiency can be remarkably improved, the risk identification capability is enhanced, the complex language understanding bottleneck is broken through, and meanwhile the manual analysis cost is reduced.

The invention discloses a video processing method and system based on a big data technology, and constructs an intelligent video processing system with semantic driving, man-machine collaboration and continuous evolution by fusing advanced technologies such as big data processing, bimodal AI analysis, knowledge graph, natural language understanding and feedback learning. Compared with a traditional video monitoring system, the scheme has the advantages that the retrieval efficiency, the semantic understanding depth, the event association analysis capability, the user interaction experience, the system self-optimization capability and the like are remarkably improved, and the problems of incomplete seeing, difficulty in finding, inaccuracy in judgment and poor use are effectively solved.

The invention discloses an advertisement generation evaluation system and method based on industrial big data, and belongs to the field of advertisement evaluation.The method comprises the steps that an advertisement information understanding model is constructed, advertisement data are imported into the advertisement information understanding model to extract advertisement semantic and visual features to recognize advertisement core focuses, and a product feature analysis model is constructed; product function parameters and selling point features are extracted through product data, product keywords are recognized, an audience cognition analysis model is constructed, audience understanding ability, information processing ability and attention focus features are extracted through user data, the language understanding and cognition level of target audiences is evaluated, and an advertisement product matching model is constructed; the matching condition of advertisement content, audiences and products is evaluated through an advertisement multi-mode semantic vector, an audience cognition vector, an advertisement focus vector and a product feature vector, a matching effect evaluation model is constructed, the defects of advertisement expression complexity, focus coverage and product adaptation degree are recognized, and the advertisement creation and delivery effect is improved.

The invention discloses a story-driven role and scene image generation method. The method comprises the steps that a natural language story text input by a user is received and preprocessed; through predefined role description structure constraints, enabling the language understanding and generation model to output a structured role description information structure under template constraints; generating a role image according to the structured role description information text, and extracting image features for consistency control; establishing a mapping table of structured role description information and image feature representation, and realizing the consistency of the appearance of roles in multiple scenes; automatically disassembling the complete story text into a plurality of scene nodes, and generating structured scene description information for each scene; and generating a complete story picture in combination with the scene description information and the role reference diagram. The invention provides a story-driven role and scene image generation method, which is used for automatically generating a story text to a role image and a scene image through semantic understanding, information description structured generation and image consistency management.

The invention belongs to the technical field of automobile testing, and particularly discloses a digital twinning-based automobile instrument testing method, which comprises the following steps of: automatically generating a structured test case through a large language model, constructing a double-digital twinning architecture by relying on a high-fidelity rendering engine and a real-time simulation engine, and realizing synchronization of a virtual model and a physical instrument; multi-modal data are collected through a multi-sensor array, and defect intelligent analysis and grade judgment are completed in combination with a large language model; and test strategy self-optimization is realized through reinforcement learning based on historical test data. According to the invention, a double-engine architecture in which a high-fidelity rendering engine and a real-time simulation engine cooperatively work is adopted, and the requirements of large-scale scene and refined rendering are met at the same time. And seamless switching and data fusion between the double engines are realized through the scene transformation matrix, so that the consistency of a virtual environment and a physical entity is ensured. Meanwhile, a large language model is introduced into a test case generation and result analysis link, and natural language understanding level test script generation and anomaly analysis are achieved.

The invention provides a multi-modal medical image intelligent analysis system and method based on a large language model, and the system comprises a key frame extraction module which carries out the key frame extraction through a multi-frame quality evaluation algorithm; the coronary artery blood vessel extraction module is used for realizing accurate extraction of a coronary artery tree structure through multi-scale blood vessel enhancement; the pathological recognition module adopts an improved DeepLabV3 + algorithm to construct a cardiac blood vessel blockage region segmentation model; the multi-modal diagnosis module adopts LLaVA-Med multi-modal diagnosis, integrates vision and language understanding ability, and combines DeepSeek to generate a standardized diagnosis report; the model optimization module is used for implementing model parameter iterative updating based on a near-end strategy optimization algorithm; and the visual interaction module is used for visually displaying the lesion position and the spatial relationship thereof, so that a doctor can conveniently carry out preoperative path planning and intraoperative navigation judgment. On the basis of a large language model and a multi-mode technology, full-process support from image uploading to diagnosis decision making is provided.

The invention discloses an Internet of Things industry intelligent customer service supervision and control system based on artificial intelligence, and belongs to the technical field of customer service supervision. Comprising an omni-channel intelligent access and intention understanding module, an AI intelligent center and complex decision module, an intelligent supervision and ethical regulation and control module, a man-machine cooperation and humanistic care module, a data-driven optimization and edge intelligent module and a security privacy and controllable treatment module, and the omni-channel intelligent access and intention understanding module is used for integrating multiple channels. Unified request distribution is achieved, data input by a user are analyzed through voice recognition, NLU natural language understanding and image recognition technologies, composite intentions are accurately captured, interaction strategies are dynamically adjusted in combination with device state data, user historical behaviors and real-time positions, and pacified talking skills or manual intervention are triggered through voiceprint / text emotion analysis. On the basis of realizing customer service supervision and regulation, all-channel fusion and intention accurate analysis can be realized, and ethical safety integrated design can be realized.

The invention provides a task processing method and device based on a large language model, electronic equipment, a storage medium and a computer program product, relates to the technical field of artificial intelligence, in particular to the technical field of large language models and natural language understanding, and can be applied to an intelligent question and answer scene. According to the specific implementation scheme, the method comprises the steps of splitting a data processing task represented by interaction data to obtain a plurality of sub-tasks; determining a plurality of large language models in one-to-one correspondence with the plurality of subtasks; and processing the plurality of sub-tasks through the plurality of large language models according to the incidence relation among the plurality of sub-tasks, and generating a task result of the data processing task. According to the method, the different sub-tasks in the same data processing task are cooperatively processed through the multiple large language models, the problem that the capacity of a single large language model is limited is avoided, and the accuracy and comprehensiveness of task results are improved.

The embodiment of the invention discloses an intention recognition method, electronic equipment and a storage medium, and the method comprises the steps: determining to-be-recognized data; the to-be-recognized data comprises at least one of text data, image data and voice data; based on the to-be-recognized data, determining a target intention element in a preset format corresponding to the to-be-recognized data through a language understanding model; and determining the target intention category corresponding to the to-be-identified data based on the target intention element, thereby realizing decoupling of natural semantic understanding and business logic, preventing a language understanding model from understanding complex business logic, and the language understanding model only needs to pay attention to extraction of the natural semantic intention element. The language understanding model is used for executing the most skilled operation of the user, and the business logic with relatively low correlation with knowledge known by the language understanding model is submitted to the post-processing process, so that the overall intention recognition accuracy can be improved.

The invention provides a method and system for pre-training a sign language understanding framework based on semantic enhancement of skeletons, and relates to the technical field of sign language recognizing.The method comprises the steps that sign language video data and skeleton sequences and text data matched with the sign language video data are obtained; extracting a skeleton key point sequence from the sign language video data, modeling to form skeleton features, and performing word segmentation processing on a text; in a pre-training stage, skeleton features and word segmentation texts are input into a fusion network to generate bidirectional enhanced features, and global and local similarities are obtained through double-layer semantic alignment; calculating the comparison loss based on the similarity, and coordinating the weight through the balance parameter to obtain the hierarchical loss; executing the matching task and the language modeling task to obtain corresponding loss, weighting and combining the three types of loss into pre-training total loss, and adjusting parameters to complete training; and finally, part of parameters are optimized in combination with specific task types in a fine tuning stage, and enhanced understanding of sign language semantics is realized. The sign language understanding accuracy is improved.

The invention discloses a healthy diet knowledge accurate retrieval engine based on a large language model. The healthy diet knowledge accurate retrieval engine comprises a user interface module, a natural language understanding module, a healthy diet knowledge base module, an accurate retrieval module, a personalized suggestion module and a feedback learning module. The engine receives healthy diet related problems through a user interface, performs natural language understanding by using a large language model, and extracts user intentions and key information. And then, related data is accurately retrieved from the healthy diet knowledge base, and personalized diet suggestions are generated in combination with the personal health information of the user. The engine also has a feedback learning function, collects user feedback and optimizes the processing logic of each module. According to the method, intelligent retrieval and personalized recommendation of healthy diet knowledge are realized, the efficiency of acquiring accurate and practical healthy diet information by the user is improved, and the user is helped to formulate and execute a more scientific and reasonable diet plan, so that the formation of a healthy lifestyle is promoted, and long-term reliable healthy diet guidance service is provided for the user.