157 results about "Word generation" patented technology

The invention relates to the technical field of natural language processing, in particular to a context-based Prompt generative model optimization system, which comprises a context analysis module, a cue word generation module, a context optimization module, a semantic check module and a structure reconstruction module. According to the method, the context path and the semantic hierarchy information of the semantic unit are introduced, the fine degree of semantic matching degree recognition is improved, semantic guide deviation caused by statement template solidification is avoided, the cue words are recombined in combination with the semantic coherence weight and the logic dependency relationship, and the recognition accuracy is improved. The consistency and expression accuracy of the prompt content in the context are enhanced, the prompt word insertion sequence and connection mode are dynamically adjusted through a semantic conflict detection and structure rechecking mechanism, coherence and stability of a semantic structure and controllable generation of the prompt content are kept, semantic conflicts and expression chaos caused by static matching are effectively avoided in the generation process, and the generation efficiency is improved. And dynamic adaptation of prompt configuration and smooth optimization of language output are integrally realized.

The invention discloses a multi-agent cooperative cue word engineering optimization method, system and device and a medium, and relates to the technical field of cue word engineering, the method comprises the following steps: after initially receiving a user input task, starting a cue word optimization process; an initial cue word template is created through a cue word generation agent, historical optimal templates are retrieved from a knowledge base for fusion, and candidate cue word templates are generated. And the analysis agent inputs the candidate template into the second large language model to generate a candidate answer. And the evaluation agent performs multi-dimensional evaluation on the answers and gives scores and defect analysis. And according to an analysis result, updating the cue word to generate a candidate template of the intelligent agent. And arranging the intelligent agent to cyclically execute the steps until the stop condition is met, and outputting the optimal cue word template. And finally, storing the optimal template into a knowledge base and establishing a semantic index to support subsequent retrieval reuse. According to the method, manual intervention is not needed, and a complete automatic process of generating, evaluating and optimizing the cue word is realized.

The invention discloses a large language model context capability excitation method and system based on an attention sink point. The method comprises the following steps: firstly, extracting a key context information vector from input data; then, in some specific Transform decoding layers of the model, anchor lexical elements in the input sequence are identified, and attention calculation paths of the anchor lexical elements are separated. In the anchor point path, contextual information is injected into representation of anchor point lexical elements through one-time cross attention calculation, and the result is stored in a key value cache for being reused in the subsequent generation step; and the non-anchor lexical elements keep the original calculation mode. And finally, the model carries out autoregression lexical element generation based on the updated sequence representation until a complete text is output. As a universal module, the method can be seamlessly integrated to any pre-training large language model based on a Transform decoder, and context information of the model is effectively excited and utilized on the premise that extra reasoning calculation overhead is hardly introduced.

The invention provides a cue word generation method, computer equipment and a computer program, and the method comprises the steps: obtaining first reply information which is generated by a target model, is used for output and corresponds to first question information, carrying out the blocking processing of the first reply information, and carrying out the caching of a plurality of pieces of first segment information; in response to second question information which is input by the user and belongs to the same dialogue scene with the first question information, retrieving target fragment information associated with the second question information in the multiple pieces of first fragment information, and based on the second question information and the target fragment information, according to a cue word template used for guiding a target model; and creating a cue word corresponding to the second question information to assist the target model in understanding target fragment information required by the context of the dialogue scene so as to generate second reply information. The processing process can be realized by a computer program based on artificial intelligence, namely an intelligent agent, and the intelligent agent can be embedded into an application degree or an operating system such as a browser or an application store to provide an intelligent function in a dialogue form.

The invention discloses a model answer generation method and device, equipment and a storage medium, and relates to the technical field of large model application, and the model answer generation method comprises the steps: obtaining a user question of a target user; inputting the user question into a user question and answer model for intention recognition, cue word generation and answer generation to obtain a model answer output by the user question and answer model; and pushing the model answer to the target user to complete user question and answer. According to the method and the device, the requirements of the user can be quickly and accurately understood, so that an accurate direction is provided for subsequent prompt word generation and answer generation, answer deviation caused by misunderstanding of the intention of the user is avoided, personalized prompts can be generated according to the questions of different users, the answers better meet the specific requirements of the user, and the user experience is improved. The pertinence and practicability of the answers are improved, so that the user question-answering model can generate higher-quality and more accurate answers, and the trust and satisfaction of the user to the question-answering system are enhanced.

The invention discloses an intelligent cue word generation evaluation method and system based on multi-component collaboration, and relates to the technical field of artificial intelligence and natural language processing. The method comprises the following steps: acquiring a multi-format text, converting the multi-format text into an input-output pair, and generating a plurality of candidate cue words; after determining a communication mode according to the text form, standardizing cue words according to the structured template of each large language model; calling each platform to generate answer analysis and unify formats, and extracting target contents by utilizing regularization; and carrying out multi-mode evaluation on the candidate cue words, selecting an optimal cue word, and outputting a structured evaluation result and an optimization suggestion based on multiple indexes. According to the method, full-process automatic processing of prompt word generation, optimization and evaluation is achieved, various model platforms, data sets, algorithms and evaluation indexes are supported, and good universality and expansibility are achieved.

The invention discloses an underground engineering on-demand ventilation knowledge graph retrieval enhancement generation method and system, and belongs to the technical field of artificial intelligence, and the method comprises the steps of data cleaning and ontology layer modeling, knowledge extraction based on prompt engineering, knowledge graph data filling and storage, storage of an extraction result in a Neo4j graph database, and retrieval enhancement generation of an inference engine. Developing an intelligent question and answer platform, and providing knowledge graph visualization and interactive question and answer functions; the system comprises a data acquisition and preprocessing layer, an ontology and knowledge extraction layer, a graph storage layer, a semantic index and entity link layer, a sub-graph arrangement and cue word generation layer, a generation and reasoning layer and an application and interaction layer. According to the invention, the acquisition and management efficiency of underground engineering ventilation knowledge is improved, the professionality and accuracy of the question-answering system are enhanced, the interface is friendly, knowledge tracing is supported, and convenient experience is provided for users.

The invention provides a dynamic interactive document generation method and device based on generative AI, computer equipment and a storage medium. The method comprises the following steps: receiving multi-modal input data; generating a composite cue word by adopting a cross-modal semantic alignment model of a QWen text encoder and a BriVL image encoder; dynamically adjusting the semantic weight of the prompt word based on a weight formula of a knowledge graph matching degree, a user historical acceptance degree and an AI confidence degree; calling the generative AI model to generate document content; a quality detection module identifies questions and actively inquires a user by adopting a layered interaction strategy; correcting the content through an incremental cue word according to the feedback; and optimizing a cue word generation strategy by adopting a PPO reinforcement learning algorithm. According to the method, the problems of static template limitation, low efficiency of cue words and lack of interactivity in traditional document generation are solved, dynamic and interactive personalized document generation is realized, and the document generation quality and efficiency are remarkably improved.

The invention discloses a Text2SQL (Structured Query Language) system and method based on cue words. The system comprises a natural language input module for receiving and preprocessing a natural language query statement; the metadata management module is used for managing database metadata information; the cue word generation module is used for generating cue words containing database modes, user problems and SQL requirements on the basis of the query statements and the metadata; the large language model module is used for generating a preliminary SQL based on the cue word; the SQL grammar verification module is used for verifying grammar and feeding back errors; the manual auditing module is used for auditing and editing the SQL statements passing the verification; and the SQL execution module is used for executing the SQL passing the audit and returning a result. According to the method, the accuracy is improved by combining cue word guidance and a large model, the field adaptability is enhanced by relying on metadata, the data security is ensured by introducing manual auditing, the accuracy and security of SQL generation in a complex query scene are remarkably improved, and the development and maintenance cost of the system is reduced.

The invention relates to a word generation method and device, electronic equipment, a non-transitory storage medium and a computer program product. The word generation method comprises the steps that a query vector Q of a previous generated word of a current word to be generated and keys K and values V of a plurality of generated words are read from a universal flash memory (UFS); based on the query vector Q and the key sum value of each generated word, calculating the importance value of the generated word; selecting a target generated word based on the importance value of each generated word; loading keys and values of the target generated word from the UFS into a memory; a word is generated using the key K and the value V of the target generated word loaded into the memory.

The invention provides a business rule cue word generation method and device, electronic equipment, a medium and a program product, relates to application of a large model in a financial science and technology scene, and can be applied to the technical field of artificial intelligence and the technical field of big data. The method comprises the steps of obtaining business demand information submitted based on a target interactive interface, and converting the business demand information into structured demand data; generating an initial cue word based on the structured demand data; inputting the initial cue word and the test data into the large model to obtain a first output result; comparing the first output result with a corresponding expected label, identifying a difference sample, and generating a cue word variant based on the difference sample; and utilizing the cue word variant to drive the large model to carry out iterative verification on the test data until the verification performance meets a preset performance condition, and outputting the service rule cue word.

The invention discloses an insurance short video content generation method and device, a medium and a program product, and relates to the technical field of artificial intelligence. Semantic analysis is performed on hot event data to extract core risk elements, so that the system can capture risk features concerned by the current society. The core risk elements are matched with the guarantee demands of the potential users, the guarantee gap is determined, and the association mapping between the hot event risks and the guarantee demands is established, so that the determination of the target insurance product has risk guidance. And determining a cue word generation strategy from a preset strategy library and inputting the cue word generation strategy into the insurance field large language model to generate a first draft of the content, thereby realizing automatic generation of the content. And in combination with a preposed compliance detection mechanism, the compliance of the generated content is ensured. Quick response and generation of the insurance short video content are realized, and the content generation efficiency is improved.

The invention provides a bridge operation and maintenance knowledge management and question answering method based on a knowledge graph and a large language model. The method comprises the following steps: performing semantic dicing and dynamic cue word generation on texts such as a design report and an overhaul report through a large language model, automatically extracting an entity-relationship-attribute triple, completing entity co-reference resolution and relationship combination, and constructing an initial knowledge graph; further utilizing a curiosity algorithm to continuously merge low-similarity new knowledge in a subsequent detection report to realize iterative growth of the atlas along with an operation and maintenance life cycle; forming a multi-level community by adopting Leiden clustering, generating an abstract pyramid, and enhancing macroscopic semantic expression; and finally, through a semantic embedding model and a GraphRAG technology, precise matching and traceable answering from natural language questions to graph knowledge points are realized. According to the method, the problems of knowledge fragmentation, update lag and illusion in the bridge operation and maintenance field are solved, and the accuracy and robustness of operation and maintenance knowledge questions and answers are improved.

The invention discloses a cue word generation method and device, equipment and a storage medium, and relates to the technical field of information processing, and the method comprises the steps: obtaining a natural language instruction and an application scene input by a target user, analyzing the natural language instruction, and extracting core intention information corresponding to the natural language instruction; matching and embedding the core intention information with a financial knowledge graph, determining related domain knowledge, dynamically fusing the core intention information, the application scene and the related domain knowledge with a multi-dimensional template, and generating an initial cue word through a cue word generation model; and performing illusion detection, fact error and logic correction on the initial cue word to generate a target cue word. According to the method, intention analysis, cue word fusion and illusion correction combined with the application scene are automatically completed in the whole process, manual intervention and optimization are not needed, the cue word generation period is shortened, and the calculation power consumption and the use cost are reduced while the cue word accuracy is improved.

The invention discloses a prompt word generation method and system based on user habits and documents, and relates to the technical field of natural language processing. According to the method, the user preference model is constructed, the initial cue word set is generated, the real-time document content is matched to screen the target cue word, the cue word is dynamically optimized, the preference model is updated through user feedback, and the cue word recommendation accuracy is effectively improved through collaborative analysis by combining the user behavior data and the document semantic features; the dynamic adjustment mechanism ensures that the system can quickly adapt to new field documents or user behavior changes, the manual annotation cost is reduced through feedback driven model updating, and finally generated cue words not only reserve long-term preference characteristics of users, but also can respond to real-time operation requirements in time, and show higher adaptability in personalized recommendation scenes.

The invention relates to the technical field of neural network models, in particular to a student model training method and device, electronic equipment and a storage medium, and the method comprises the steps: obtaining an algorithm optimization model output generated by an algorithm optimization student model based on at least one question, obtaining a basic reward based on the algorithm optimization model output and a preset answer, obtaining an answer length reward based on the length output by the algorithm optimization model and an answer length threshold; obtaining a first word generation probability that next words with different contents in the output of the algorithm optimization model are generated into different words, and obtaining a first generation entropy based on the first word generation probability; training the algorithm optimization student model based on the answer length reward and the first generation entropy to obtain a calculation reward optimization student model; and if the calculation reward optimization student model is converged, outputting. According to the scheme, it can be ensured that the calculation reward optimization student model generates correct and brief answers, the convergence speed of the model in the training process is increased, and the user experience is improved.

The invention discloses a power industry cue word optimization method and system, and the method comprises the steps: carrying out the static structured optimization of a non-standardized cue word inputted by a user, obtaining a structured cue word which accords with the AI model understanding logic and the power industry specification, and storing a data record generated by the static structured optimization into a static optimization result temporary storage library; constructing a reinforcement learning model, carrying out offline training and online iteration, carrying out secondary optimization on the structured cue word by utilizing the trained reinforcement learning model, and inputting an optimization result into an AI model; and storing an output result of the AI model, user feedback, a data record generated by static structured optimization of the cue word and data generated by secondary optimization structured cue word of the reinforcement learning model into a cue word optimization case library. According to the method, the accuracy, the reliability and the adaptivity of the cue word generation result are improved, the application of a large model in the power industry is further promoted, and powerful technical support is provided for intelligent transformation of the power industry.

The invention provides a cue word generation method and computer equipment, and the method comprises the steps: decomposing task description information in response to the task description information input by a user, and determining task elements of a target task corresponding to the task description information; the task elements at least comprise a task purpose, an execution role for executing a target task and execution steps required for completing the target task; determining a target matching degree between the task target and the task description information; in response to the condition that the target matching degree meets the cue word generation condition, determining a quality evaluation cue word for the task cue word of the target task based on the task element of the target task; and inputting the quality evaluation cue word into at least one first model to guide the corresponding first model to perform quality evaluation on the task cue word, and outputting the task cue word with qualified quality.

The invention provides an adaptive test case generation method and system based on a large model, and the method comprises the steps: calling a pre-constructed cue word generation agent in response to an instruction for generating a test case, and generating a matched target cue word and a recommended target test case generation model based on data input by a user; generating a first test case based on the target cue word by using the target test case generation model; displaying the first test case, and obtaining a first label input by a user for the displayed first test case; the target test case generation model is called, a second test case is generated in combination with the user intention and the first test case, and the matching degree of the second test case and the test case expected by the user is higher than the matching degree of the first test case and the test case expected by the user.

The invention discloses a dynamic news abstract generation method based on voice and visual perception, which comprises the steps of voice recognition text extraction, audio-text alignment, prompt word generation and final news abstract generation. The objective is achieved through an innovative process comprising a plurality of key optimization links. A robust long audio processing and recognition mechanism, an optimized audio-text alignment method, an elaborately designed cue word framework and an efficient fine-tuning multi-modal large language model are integrated. According to the method, deep understanding of multi-modal information and high-quality text generation are jointly realized, and the quality of four core dimensions including content coverage, semantic accuracy, logic continuity and language fluency of the abstract is comprehensively improved.

The invention provides an SQL cue word generation method and device and electronic device.The method comprises the steps that a data set construction step is conducted, specifically, a field data set and an SQL data set are established, the field data set comprises standard information, and the SQL data set comprises SQL statements; an information identification step: identifying key information in a query natural language input by a user; a query natural language standardization step: respectively determining first standard information, second standard information and third standard information corresponding to the key information based on the domain data set; a standard term determination step: determining a standard term for querying the natural language based on the first standard information, the second standard information and the third standard information; an SQL statement generation step: generating an SQL statement corresponding to the standard term; and a prompt word generation step: determining a result natural language corresponding to the SQL statement, and outputting the result natural language data to the user. The problem that in the prior art, the query efficiency of NL2SQL in a specific field is low is solved.

This disclosure describes methods, non-transitory computer readable storage media, and systems that extract a definition for a term from a source document by utilizing a single machine-learning framework to classify a word sequence from the source document as including a term definition and to label words from the word sequence. To illustrate, the disclosed system can receive a source document including a word sequence arranged in one or more sentences. The disclosed systems can utilize a machine-learning model to classify the word sequence as comprising a definition for a term and generate labels for the words from the word sequence corresponding to the term and the definition. Based on classifying the word sequence and the generated labels, the disclosed system can extract the definition for the term from the source document.

The invention discloses a visual analysis system for harmful cue words based on risk perception. The system comprises a sidebar module at the front end, which is used for importing models and data and displaying theme distribution; the knowledge graph exploration module is used for displaying the coverage degree of risk concepts and revealing a risk coverage gap; the harmful prompt word view browsing module is used for reviewing and editing harmful prompt words; the strategy adjustment panel module is used for adjusting a generation strategy based on the performance condition of data synthesis; the index trend view module and the index monitoring view module are used for helping a user to pay attention to a harmful cue word performance expression fluctuation root and evaluating a harmful cue word iteration effect through indexes respectively; and the version switching module is used for multi-dimensional version comparison. And constructing a rear-end calculation module, and driving knowledge graph construction, harmful cue word generation and multi-dimensional index calculation according to front-end configuration to iteratively optimize the harmful cue words. Through front-end and back-end man-machine cooperative control, a field security enhancement closed loop of data synthesis, vulnerability positioning, strategy regulation and control and iteration enhancement is realized.

The application discloses a method and system for generating collusion attack samples of peer review system. The application mainly aims at the problem that the peer review system colludes with the text submitter and the reviewer to construct an antagonistic text to manipulate the results of the peer review distribution system, explores the method of collusion attack in the peer review scene, and the method comprises the following steps: collecting target text and the text of the colluder in the participating reviewer and performing text preprocessing; finding a target attack word; constructing a candidate attack word table; replacing words according to the order of the attack word table to generate an antagonistic sample. The application replaces the synonyms and the near synonyms of the target text and the colluder text, makes the least changes to the sentences of the target text, generates an antagonistic sample that cannot arouse human awareness in the maximum extent, reduces the quality of the peer review distribution, improves the success rate of the antagonistic sample, and reduces the time and complexity of the algorithm.

Embodiments of the present application provide a prompt word generation method and device, equipment, storage medium and program product, relating to the technical field of financial technology and artificial intelligence. The method comprises: collecting first interaction behavior data between a user and a large language model and generating a first interaction feature vector based on the first interaction behavior data; inputting the first interaction feature vector into a first model to obtain a quality score of a first prompt word of the large language model; in the case that the quality score of the first prompt word is lower than a first threshold, generating a rewrite trigger signal and generating a new version of the first prompt word based on the rewrite trigger signal. The method of the present application can improve the efficiency of prompt word generation and the performance of the large language model.

The invention discloses a query word generation method and device, electronic equipment and a storage medium, and relates to the field of computers, in particular to the technical field of artificial intelligence such as large models, deep learning and intelligent recommendation. According to the specific implementation scheme, firstly, a query request is received, a search result page is displayed based on a query word, then, a first detail page associated with first content is displayed under the condition that the first content in the search result page is clicked, then, a first recommendation query word is generated based on first associated information of the first content and the query word, and the first recommendation query word is recommended to the search result page. And finally, updating the search result page based on the first recommendation query word to obtain an updated search result page.

The invention discloses a method and system for integrating retrieval and enhancing generation ability, and the method comprises the steps: S1, building a basic SQL set, and enabling each basic SQL to correspond to a preset business intention and to comprise a placeholder condition field; s2, establishing a mapping relation between fields and noun types; s3, performing intention analysis by using a large language model, and matching a basic SQL; s4, performing query condition reasoning, extracting specific nouns, and executing expansion word generation and compound word processing; s5, generating a complete SQL according to rule driving; s6, executing the SQL and dynamically correcting a query result to enable the query result to meet a context length requirement; and S7, inputting the question and result combination into the large language model to generate an answer. The method does not need a vector database, adapts to a small parameter model, is low in cost and high in real-time performance, and can be widely applied to various traditional service systems needing to integrate the AI dialogue ability.

The invention discloses a field large model harmful cue word generation method based on a knowledge graph. The method comprises the following steps: screening a universal harmful cue word data set based on a constructed risk knowledge graph to obtain a seed harmful cue word library; processing the field corpus to obtain an embedded context; generating candidate harmful cue words through a synthesis model based on seed harmful cue words, risk entities, embedded contexts and examples, cleaning and enhancing toxicity indexes to obtain high-risk harmful cue words, and adding the high-risk harmful cue words into a seed harmful cue word library; and based on the semantic relevancy between the seed harmful cue words and the risk entities and the toxicity scores of the seed harmful cue words, screening out cue word input of a next round, and constructing an iteratively updated field harmful cue word data set. The method has the characteristics of high automation, multi-dimensional evaluation, controllable generation and the like, multi-round man-machine collaborative prompt word construction can be realized, and the red team drilling efficiency and the safety test quality of a large language model in a specific application field are remarkably improved.

The invention discloses a real-time context sensing dynamic cue word generation method based on large model translation. The method comprises the steps of initialization, context sensing and core concept extraction, dynamic cue word generation, new concept recognition and list updating. According to the method, a traditional static translation process is converted into a dynamic process with memory and state maintenance capabilities, a global core concept list is constructed and dynamically maintained in real time, and dynamic cue words containing determined standard translations are automatically generated in the translation process; the global consistency of the translations on core concepts such as term named entities is remarkably improved, the problem that core concept translations are inconsistent in long document translation is fundamentally solved, and the method is suitable for professional fields such as technical documents, academic papers and commercial reports with extremely high term consistency requirements.

This application provides an event detection method and apparatus based on a graph perturbation strategy, including: acquiring context words in a given sentence; generating a syntactic information adjacency matrix and a concatenated vector corresponding to the context words; generating a syntactic graph based on the syntactic information adjacency matrix; determining a context representation corresponding to the context words based on the concatenated vector; using the syntactic information adjacency matrix and the context representation as input to an artificial neural network to obtain output information; and predicting the type of the context words based on the context representation and the output information. This application effectively solves the problem of low classification efficiency caused by excessive redundant information in long sentences during event detection by introducing syntactic information and two graph perturbation strategies to filter redundant information in sentences, retaining important word information, and using graph repair operations to reduce the loss of syntactic information.