156 results about "Specific knowledge" patented technology

The invention provides a large model cue word design method, system and device in an industrial scene and a medium, and belongs to the technical field of artificial intelligence. The method comprises the following steps: constructing a private cue word template library, and forming a domain special knowledge unit set through multi-source industrial data collection, intelligent semantic analysis and classified storage; customizing a four-layer structured template, sequentially establishing a task overview layer, a step disassembly layer, an instruction refining layer and an output specification layer, and constructing a layered mapping model from task definition to output execution; template library management and iterative optimization are carried out, and dynamic updating and performance improvement of a cue word system are realized through centralized system management, multi-source feedback acquisition and data-driven optimization. According to the method, a cue word design and optimization system based on a hierarchical structure is constructed for complex task requirements in an industrial scene, and the task processing performance of a large model in scenes of industrial production, quality detection, equipment operation and maintenance and the like is improved.

The invention discloses a knowledge and data fused chronic disease collaborative management method and system, and relates to the technical field of medical management. Comprising the following steps: generating a feature data set, processing the feature data set by using a feature gain modeling method, and generating an enhanced feature data set; dynamically generating a chronic disease comprehensive knowledge graph, performing node supplementation and weight optimization on the knowledge graph, and generating a specific knowledge sub-graph of the patient; carrying out multi-time sequence prediction by utilizing the specific knowledge sub-graph, and carrying out short-term and long-term trend prediction; a patient health track model is constructed through the state transition probability in combination with the prediction result; according to the dynamic risk labeling result of the health track model, the health state of the patient is comprehensively evaluated in combination with the specific knowledge sub-graph, and personalized management suggestions are generated according to the evaluation result. According to the collaborative management platform, cooperation between different medical institutions and medical staff can be promoted, information sharing and resource integration are realized, and the overall level of medical services is improved.

The invention relates to the technical field of multi-round dialogue processing, in particular to a multi-round dialogue processing method based on RAG technology dynamic cue word retrieval. Comprising the following steps of question input by a user, initial retrieval and cue word generation, question splicing and answer generation, historical dialogue management and context coding, multi-round question analysis and retrieval enhancement, complex question decomposition and parallel retrieval, multi-source information fusion and generation model optimization, iterative feedback and model tuning, and exception handling and fault-tolerant mechanism. According to the method, related information is retrieved from an external knowledge base, the model can utilize wider and more specific knowledge to generate answers and better track dialogue history and context, so that answers consistent and coherent with the dialogue history are generated, the model can generate answers on a short input sequence by retrieving and splicing related cue words, and meanwhile, the model can be used for analyzing the dialogue history. Cue words are dynamically generated according to user input and retrieval results, so that the model can better adapt to different dialogue scenes and user requirements.

Systems, methods, and computer-readable media provide a context-specific prompt to answer a user query. The systems, methods, and computer-readable media determine a context based on content of a natural language request and / or determine a role of a user who submitted the natural language request. Additionally or alternatively, templates or RAG sources that will be used for prompt generation may include financials domain-specific knowledge or other domain-specific knowledge or insights. Inclusion of this additional information in the prompt enhances the context to promote more accurate results from a large language model. In one embodiment, the prompt templates are created from various RAG sources, such as payables, general ledger, receivables, and asset management, containing structured data and information specific to the financial domain, enterprise, or other domain, which helps craft accurate prompts. A prompt is generated that identifies a subset of available fields and other selected information based on the role or other context. The prompt template may contain domain-specific knowledge uses a relevant domain or enterprise information to drive relevant results, and an executable query is generated by a large language model based on the prompt. The executable query causes data to be retrieved from a database to generate a result, and information is displayed based at least in part on the result.

Computing units provided at local sites or edge locations are programmed to execute conversational tools that generate pertinent, domain-specific responses to queries received from workers at such sites or locations. The conversational tools are large language models that are trained with domain-specific knowledge documents. Data representing queries are received from workers at such sites or locations and provided as inputs to the conversational tools along with text representing nearest knowledge documents from a knowledge base associated with the domain, as well as contextual data. Responses identified based on outputs received from the conversational tools in response to the inputs are provided to the workers that generated the queries. Where subsequent queries are received from the workers, responses to the subsequent queries are identified based on the subsequent queries, nearest knowledge documents, contextual data, and conversational histories including previously received queries and responses to such queries.

The invention relates to the technical field of artificial intelligence, and particularly discloses a mutual learning method and system between a multi-view medical image and a text report, and the method comprises the steps: constructing a mutual learning model, and dividing the learning process of the mutual learning model into an image reconstruction task, a report reconstruction task and a multi-view alignment task; the image reconstruction task obtains the feature representation of a multi-view medical image, the report reconstruction task obtains the feature representation of a text report and performs cross-modal fusion on the feature representation of the text report and the image feature representation to obtain a reconstructed text report, and the integration of the multi-modal reconstruction tasks enables the model to learn richer and more detailed feature representation. And the robustness to lost or damaged data is improved. In addition, the exchange-based multi-modal fusion method in cross-modal text reconstruction aims to fully integrate visual features and enrich semantic representation of specific knowledge in the field. Through pre-training and joint training, the model performance is optimized, so that a more accurate and comprehensive diagnosis result can be obtained after the mutual learning model is applied.

Automated computer systems and methods to determine a sentiment of information in digital information or content are disclosed. One aspect includes deriving, by a processor, the digital information from a source; generating, by the processor, a domain-specific machine learning sentiment score, based on the digital information, by one model of at least two machine learning models; autonomously mapping, by the processor, a non-domain specific knowledge graph of associations between elements in a set of digital contextual information; receiving, by the processor, sentiment graphs, each sentiment graph defining a sentiment; generating, by the processor, a graph sentiment score based on the non-domain specific knowledge graph and the sentiment graphs; generating, by the processor, a final sentiment score based on the graph sentiment score and the domain-specific machine learning sentiment score; and determining the sentiment of the information in the digital information or content via the final sentiment score.

The invention provides a complex scene-oriented end-to-end semantic extraction system, belongs to the technical field of artificial intelligence and natural language processing, and realizes cross-modal information association through a multi-source heterogeneous data fusion module to construct a dynamic semantic network model. A hierarchical attention mechanism is adopted to carry out context-aware coding on unstructured input, and unsupervised pre-training and a weak supervised fine tuning strategy are combined to optimize a feature representation space. And designing an adaptive inference engine, automatically switching semantic analysis paths based on scene complexity, and generating a structured output result. According to the method, the dependency on specific knowledge in the field is reduced, the semantic understanding generalization ability in a complex scene is remarkably improved, high-precision analysis performance can still be kept in a low-resource environment, meanwhile, calculation resource consumption is reduced, and the method is suitable for practical application scenes with multi-language mixing, serious noise interference and high real-time performance requirements.

The invention relates to the technical field of natural language processing, and discloses an SQL code generation method and device, and a financial data query method, device and equipment, and the method comprises the steps: obtaining a question statement sent by a target object for a target field; performing intention recognition on the question statement by utilizing the generated large model so as to construct a context cue word; wherein the generated large model is obtained by training generalization knowledge and specific knowledge in advance; performing accurate matching of an SQL generation template on the question statement according to the context cue word; if the accurate matching result shows that the target SQL generation template is matched, generating an SQL code corresponding to the question statement based on the target SQL generation template; and if the accurate matching result shows that the target SQL generation template is not matched, generating an SQL code corresponding to the question statement according to the context cue word. The SQL code generation method and device have the beneficial effects that the efficiency and accuracy of the SQL code generation process are improved, and meanwhile, the matching degree of the obtained SQL code and the question statement is improved.

The embodiment of the invention discloses a knowledge question answering method and device, equipment, a medium and a program product. The method comprises the following steps: displaying at least one first tag in a knowledge question-answer interface; receiving a target label selected for the question information; and outputting answer information corresponding to the question information in the knowledge question and answer interface. According to the embodiment of the invention, the specific knowledge content in the knowledge base is indexed through the tag, so that the answer personalization is improved and the knowledge question-answering efficiency is improved.

A large-scale language model for adapting to an industrial domain is provided, a graph database stores a domain-specific knowledge graph containing instance-level semantic information about the industrial domain, including true triple statements about physical entities from the industrial domain and their interrelation. A knowledge graph to corpus translator converts the domain-specific knowledge graph into a natural language corpus. A domain adaptation component pre-trains the large-scale language model with the natural language corpus to a provide a domain-adapted large-scale language model. This approach does not require any assumptions on the structure of the domain-specific knowledge graph, for example, it does not require the domain-specific knowledge graph to contain only instance data. Neither does it need linguistic templates since the verbalization of triples can be built on their existing labels or relation names. The triple-based information encoding of RDF facilitates a natural translation into language during the converting operation.

The invention relates to the field of artificial intelligence, the technical scheme is applied to the financial field and the medical health field, and discloses a federated learning-based large model optimization method and device, computer equipment and a storage medium, and the method comprises the steps of receiving a global model from a server; performing local training on the global model through local data, keeping weight freezing of the basic model during local training, and updating the first low-rank matrix and the second low-rank matrix; sending the updated first low-rank matrix to a server, so that the server aggregates the updated first low-rank matrix of each client to obtain a global low-rank matrix; and receiving the global low-rank matrix from the server, generating a weight update quantity to update the weight of the local adapter of the global model, and obtaining an optimized global model. According to the scheme, the communication efficiency can jump; the second low-rank matrix solidifies client-side specific knowledge to prevent global knowledge pollution; and a weight merging mechanism in the reasoning stage ensures zero delay spread.

Techniques for constraining the results of a generative language model to valid information using knowledge-grounded documentation. A generative language model may generate invalid results, including compound entities and incorrect entity relations. The techniques include, for a given user inquiry, determining a set of documented information, from a particular knowledge base, that corresponds to the user inquiry. The techniques further include determining a subgraph from a knowledge graph representing the knowledge base, as well as determining a trie data structure representation of the set of documented information. The user inquiry and subgraph are provided as input to a trained generative language model for generating a response to the user inquiry. The techniques include using the trie data structure to validate that the generated response corresponds to real information from the set of documented information.

The invention discloses a method for generating knowledge answers to a mobile phone retail store based on LIama3 and retrieval enhancement. The method comprises the following steps: collecting related information of a sales scene of the mobile phone retail store and performing text processing to form a data set; converting texts in the data set into vectors through a BGE model, and storing the vectors into a local vector knowledge base; recognizing a voice problem uploaded by a user by using the finely-adjusted Whisper model, and converting the voice problem into a problem vector; documents related to problem vectors are retrieved in a local vector knowledge base through RH-FAISS Merge mixed search, and a final knowledge document is selected through a Rerank method; generating cue words by using a cue word template Prompt according to the final knowledge document; and submitting the generated prompt word to the Llama3 model to generate an answer, and displaying the answer to the user. According to the method, the problem that related problems of mobile phone retail cannot be accurately understood and answered due to the fact that an existing large language model lacks enough specific domain knowledge during training is solved.

The invention belongs to the technical field of artificial intelligence, and particularly relates to a security detection method and system based on a large language model and retrieval enhancement, and the system comprises an IDE plug-in module and a security language server SLS module; the IDE plug-in module is used as a front end for interaction between a user and a system, is directly integrated in a development environment IDE, and communicates with a security language server SLS module at a rear end by using an API provided by the IDE to realize real-time capture of code context and visual presentation of an analysis result; the method has the advantages that the strong code understanding ability of the LLM and the specific knowledge of the field provided by the RAG are combined, the knowledge limitation of the universal LLM in the professional security field is effectively relieved, and the detection accuracy and the relevance of suggestions are improved. The risk can be detected, detailed explanation and specific repair suggestions can be provided, and the understanding and repair cost of developers is reduced.

The invention relates to the field of power grids and semantic communication, and particularly discloses an intelligent power grid defect data acquisition and classification method based on semantic communication, which is characterized in that a large-scale model is used for extracting specific knowledge in the field of intelligent power grids, a vector quantization compression mechanism is used for efficient transmission, and a two-stage training strategy is used for optimizing a training process. The method comprises the following steps: constructing a semantic encoder based on a large-scale model; carrying out vector quantization on the image features, and carrying out a de-quantization process on a quantization result at a receiving end by simulating a channel of an actual scene; decoding operation is carried out on the de-quantization result, and a classification task of the defect image is completed; performing two-stage training on the system to obtain a codec with optimal performance and a vector quantization module; the transmission and classification of inspection defect pictures are completed by using the trained system, and the classification accuracy and the transmission data volume are better.

The invention relates to the technical field of code generation, in particular to a knowledge-adaptive code retrieval model, method and system. The knowledge adaptive code retrieval model provided by the invention comprises a code language selection module, a code knowledge module, a general knowledge module and a feature fusion module, the code language selection module and the feature fusion module are respectively connected with the code knowledge module and the general knowledge module; the code language selection module selects appropriate code languages for different user questions, the various code knowledge modules learn specific knowledge of each code language, and the universal knowledge module captures user intentions or universal knowledge of various code languages, so that under the interaction of the universal knowledge module and the various code knowledge modules, the user questions can be learned through the code language selection module. And the learned feature embedding is enabled to be richer and more professional, so that the accuracy of a code retrieval task is improved. And the code language selection module adaptively selects a proper code language for the user problem, and considers the applicability and the applicability degree of each code language, so that the code retrieval accuracy can be improved.

The invention discloses an interpretable clinical decision support system based on label generation and a knowledge graph. The method comprises the following steps: based on a breast cancer domain knowledge enhanced version Qwen-BrCaAdapt of a general large language model Qwen, analyzing an unstructured medical record text of a patient, and generating a structured result containing tags, values, evidences and explanations; calculating a reasoning label through a path matching engine by utilizing an editable structured path rule table, and matching a candidate treatment scheme according to the reasoning label; and taking the matched treatment scheme as a central node, calling a medical knowledge graph to bind entity information including clinical evidence, recommendation levels, medical insurance information, medication risks, usage and dosage, and generating a traceable JSON structure and a visual report. The method has the beneficial effects that the accuracy and efficiency of tag generation in the breast cancer field are improved, the rule maintenance cost is reduced, the interpretability and traceability of a clinical decision scheme are enhanced, and the acceptability of a doctor to a recommendation result is improved.

The invention discloses a multi-modal rumor detection system based on multi-field perception, relates to the technical field of rumor detection, and is used for solving the technical problem of low multi-field rumor detection precision in the prior art. The rumor detection system comprises a multi-view feature extraction module, a multi-domain perception module, a multi-level knowledge fusion module and a final classifier. Wherein the multi-view feature extraction module is used for respectively extracting corresponding single-mode features from a text mode and an image mode, and acquiring aligned image and text features as multi-mode view features; the multi-domain sensing module is used for learning cross-domain shared knowledge and domain specific knowledge at a text view angle, an image view angle and a multi-mode view angle; the multi-level knowledge fusion module is used for fusing cross-domain shared knowledge and domain specific knowledge of a text view angle, an image view angle and a multi-mode view angle; and the final classifier is used for obtaining a prediction result according to the cross-domain shared knowledge and the domain specific knowledge obtained by the multi-level knowledge fusion module.

The invention belongs to the technical field of artificial intelligence, and particularly relates to an enterprise special agent rapid construction method based on large model fine tuning. The method comprises the following steps: S1, performing agent basic configuration; s2, outputting an enterprise special agent based on the agent basic configuration scheme and the general large model; s3, setting rule parameters of an agent response strategy, and generating an agent interaction strategy; s4, performing full-life-cycle management on enterprise knowledge entries to ensure the accuracy and traceability of agent answering; and S5, verifying the accuracy of the agent knowledge and continuously optimizing the agent knowledge. According to the invention, through low-code configuration, an enterprise can construct an agent without a technical team, and autonomous controllability is realized; a large model adaptation technology enables the intelligent agent to accurately understand proprietary knowledge, solves the problem of term misunderstanding, and achieves high-precision adaptation; the knowledge full-life-cycle management and rapid updating mechanism ensures the timeliness of the agent knowledge; customizable interaction rules enable the intelligent agent to adapt to multi-scene requirements, and the user experience is improved.

Context-aware content filtering adapted for a knowledge domain is provided. In certain examples, a classification confidence score by a classifier indicates a level of confidence that a prompt from a user is associated with the knowledge domain. The classification confidence score is compared with a threshold. When the score is below a threshold, a violation notice is provided to the user without submitting the prompt to a generative artificial intelligence (GAI) model. When the classification confidence score is above the threshold, the prompt is further processed to determine, according to rules, whether the prompt should be submitted to the GAI model. In various examples, the rules are applied to contextual information, safety score information, and intent information derived from the prompt.

The present invention provides a large language model prompt word engineering method and system based on multi-round RAG technology, belonging to the field of artificial intelligence large language models. The present invention constructs a group of prompt words including background knowledge, industry-specific knowledge, and preliminary reasoning results, and inputs the group of prompt words into a large model with a low training parameter scale and no industry-specific knowledge fine-tuning training, so as to significantly improve the knowledge scope, recall rate (Recall), precision rate (Precision) and F1 score (F1 Score) covered by the reasoning results, thereby achieving the same reasoning results. Under the premise of using a large model with a lower parameter scale and less computing resource consumption, the cost of using the large model is reduced; the method and system include a prompt word construction step and a large model question-answering system with low computing power consumption, and the large model question-answering system with low computing power consumption includes: a pre-trained large language model, question-answering system application software and a low-cost general-purpose personal computer running these programs and software.

The embodiments of the present application provide a question-answering model training method, device, equipment and storage medium. The method includes: obtaining historical customer service information; constructing a thinking chain in the customer service processing process based on the historical customer service information; based on the thinking chain, selecting multiple customer service data from the data set for annotation to obtain annotated data; training the initial model based on the annotated data to obtain a trained question-answering model. The present application extracts key information from historical customer service information, and constructs a thinking chain in the customer service processing process based on this key information, and integrates the thinking chain into the training data of the model, ensuring that the training data matches the logical structure of the thinking chain, which can effectively guide the learning process of the model, deeply understand the customer service processing business rules, and improve the accuracy of the model's response; in addition, integrating the thinking chain into the training data of the model also allows the present application to be independent of a specific knowledge base, simplifying the deployment and maintenance costs of the model.

Collecting task-specific knowledge for a classification task by AI in the industrial environment requires a high degree of expert knowledge about the respective area. For the purpose of formalisation and embedment in the machine learning model used, ML experts are however required. Close cooperation between data scientists and domain experts for each specific application is a prerequisite. This process is very costly and prone to errors in the event of incorrect communication, and as a result, these approaches are practical only for a limited number of applications where the benefit of solving the problem with an ML approach with a high degree of domain knowledge integration is sufficiently high to cover this cost. The proposed method overcomes the above-mentioned hurdles by providing a way of using knowledge contained in LLMs in small application-specific models. This is done by distilling expert knowledge, which is combined into rules for the direct classification and / or extraction of relevant data features in LLMs using domain-specific input requests and is then used in small task-specific models during inference or training.

The present application relates to a translation system for natural medicinal material domain-specific knowledge. The system comprises: a front-end program, a first learning model, a search engine and a natural medicinal material domain-specific knowledge base. On the basis of performing coreference primary term annotation, and coreference annotation of standard translations in multiple languages on natural medicinal material terms in the natural medicinal material domain-specific knowledge base, the search engine is used to search in the natural medicinal material domain-specific knowledge base for a natural medicinal material term extracted by the first learning model from text to be translated, the first learning model generates a translated document in an original format on the basis of a search result, and finally, the front-end program displays a translation result in a target language in a mode matching a display mode of a user interaction interface. By means of the translation system in the present application, a user can conveniently and intelligently perform accurate and standardized translation on related terms and knowledge in the domain of natural medicinal materials on the basis of requirements of the user.

The invention provides a heterogeneous knowledge migration method and system based on redundancy suppression, and the method comprises the steps: respectively sending an input image into a heterogeneous teacher network and a heterogeneous student network, extracting the feature vectors of the networks, and obtaining a teacher feature vector and a first student feature vector; the first student feature vector is sent to a decoupling module, a second student feature vector consistent with the teacher feature vector in dimension is obtained, the first student feature vector is decoupled from a redundancy suppression optimization target, and specific knowledge of a student network is reserved; calculating redundancy suppression distillation loss based on the teacher feature vector and the second student feature vector; calculating cross entropy loss based on the first student feature vector and a truth value label of the input image; and summing the redundancy suppression distillation loss and the cross entropy loss as total loss, and training a student network and a decoupling module. According to the invention, the student network can avoid learning specific incompatible information of a teacher network architecture and only focuses on knowledge beneficial to the student network, so that the heterogeneous distillation effect is improved.

The invention discloses a robot autonomous assembly method and system based on imitation learning, a storage medium and computer equipment, and the method comprises the steps: constructing a universal expert strategy suitable for polygonal shaft hole assembly through analyzing and summarizing the polygonal shaft hole assembly experience of an expert; according to the strategy, task type images and human assembly demonstration data including assembly force / torque and assembly actions are collected; modeling a human assembly skill into a universal assembly skill model based on shared knowledge and specific knowledge; using a loss function guidance algorithm to train and optimize the general assembly model; and deploying the optimized general assembly model into a control system of the robot so as to drive the robot to execute humanization-like assembly. According to the method, the problems of poor generalization performance, slow convergence of an end-to-end learning mode and high cost of an imitation learning method can be solved, rapid learning of human skills can be realized, and humanoid robot autonomous assembly is realized.

The invention discloses an epilepsy detection model training method, an epilepsy detection method and an epilepsy detection system, and belongs to the technical field of epilepsy detection. When a domain training subset S of a new patient is obtained, a new training batch is started, an epilepsy detection model is trained based on the domain training subset S under the current training batch, a sub-network is distributed for each domain through adaptive pruning, specific domain parameters are combined to learn specific knowledge of the new domain, and the epilepsy detection model is obtained. And meanwhile, the shared parameters effectively utilize knowledge learned in a previous domain. Self-adaptive pruning promotes knowledge sharing, reduces inter-domain interference to the greatest extent, effectively relieves disastrous forgetting, and can realize accurate epilepsy detection; meanwhile, data of a previous domain does not need to be stored, and data privacy is effectively protected; on the basis, on the premise that patient data privacy is protected, accurate epilepsy detection is achieved.

The invention discloses a urology department traditional Chinese medicine inquiry result prediction method and system applying a large language model, and belongs to the technical field of large language models. The method comprises the following steps: preprocessing disease description data, and obtaining self-correlation parameters among the disease description data; constructing a prediction model according to the large language model; the core keyword data used for describing the focus is output, so that the specific cause of disease in the inquiry is predicted, and the accuracy is improved; the method comprises the following steps: recording feedback data related to core keyword data in real time; inputting the feedback data into the self-correction model, and outputting an adjustment value; and according to the adjustment value, parameters in the large language model are adjusted in real time, and through continuous use and adjustment based on feedback data, continuous refinement of the prediction model in the specific knowledge field of the urology department field is further realized, so that prediction of specific causes in inquiry is further realized, and the accuracy is improved.

The present disclosure relates to a technique for addressing an issue to be resolved associated with an electronic document. The method discloses accessing an actionable portion associated with a particular knowledge domain of the electronic document and associated context. Further, retrieve data from data sources to provide additional information related to the particular knowledge domain and the associated context. Then structuring the retrieved data to produce a subset of organized data and determine the issue to be resolved related to the electronic document. Further, generate data elements associated with the issue to be resolved and map dependency relationships between data elements. Also, determine test goals associated with the issue to be resolved based on the dependency relationships. Thereafter, determine corresponding test cases associated with resolution and determines actionable test steps related to the issue to be resolved based on the corresponding test cases associated with the electronic document.