689 results about "Complex problems" patented technology

The invention discloses a knowledge exploration method and system based on a generative thinking chain and a feedback mechanism, and the method comprises the steps: receiving a multi-modal query inputted by a user, and carrying out the intention analysis to recognize a core demand; preprocessing the multi-modal query, extracting key information, and decomposing a complex problem into a plurality of sub-problems or sub-tasks; dynamically constructing a hierarchical reasoning path based on the generative thinking chain, and marking a logic basis for each step of reasoning path; reasoning path priority is optimized through dynamic weight distribution, a reasoning strategy is dynamically adjusted on line in combination with a meta-reinforcement learning framework, and a reasoning path is optimized; verifying the reasoning path by using a causal reasoning analysis tool, and generating an anti-fact path to correct error nodes; and outputting a structured text and a visual reasoning process, and optimizing model parameters in real time based on multi-user feedback aggregation of the game theory. According to the method, the problems of intelligent query and reasoning in a complex scene can be effectively solved, and the accuracy, real-time performance and credibility of knowledge service are realized.

The invention relates to the technical field of digital factory production management, and discloses a digital factory-oriented multi-target production plan optimization method and system, and the method comprises the steps: decomposing an enterprise order into a plurality of subtasks through a production task decomposition module, and extracting related information; the resource dynamic evaluation module collects data in real time to generate a resource state matrix; the multi-objective optimization algorithm module constructs a model and adopts an improved non-dominated sorting genetic algorithm to solve; the dynamic priority distribution module adjusts the priorities of the sub-tasks according to the real-time data; and the conflict resolution unit is used for solving resource allocation conflicts. The method can accurately decompose tasks, evaluate resources in real time, collaboratively optimize multiple targets, flexibly deal with abnormities, efficiently resolve conflicts, comprehensively improve the production efficiency and benefits of the digital factory, and effectively solve the complex problems in the production plan making and optimizing process of the digital factory.

The invention discloses a privacy enhanced intelligent search method and system based on multi-round iteration. The method comprises the following steps: performing hierarchical semantic analysis on a query input by a user; splitting the complex query into sub-queries based on a task dependency graph algorithm; according to the sub-query, retrieving an evidence fragment from the multi-source data, constructing a semantic element coverage matrix to detect a knowledge gap, and if an uncovered element exists, generating a supplementary sub-query for iterative completion until a preset termination condition is met; integrating cross-modal data through a federated learning technology, and generating a structured knowledge graph fragment in combination with semantic vector alignment and an evidence fusion algorithm; performing dynamic desensitization processing on the retrieval result; and a closed-loop iterative updating mechanism is formed based on a user explicit and implicit feedback optimization retrieval strategy. The problems of traditional intelligent search in the aspects of semantic understanding depth, complex problem reasoning, search result accuracy and integrity, user privacy security and the like are effectively solved.

The invention discloses a medical information multi-agent expert thinking chain collaborative reasoning method and system, and the method comprises the steps: constructing a sub-expert agent and a summarizer agent based on a large language model, processing the medical information, and outputting a processing conclusion of the medical information; and secondly, constructing an evaluator agent, performing medical information conflict evaluation on the output of the sub-expert agent and the summarizer agent, activating a multi-agent debate mechanism when the conclusion output of the summarizer agent exceeds a medical conflict threshold, and outputting whether a processing conclusion is credible or not. Then constructing a coordinator agent, selecting called sub-expert agents and summarizer agents, and generating a plurality of forward reasoning thinking chains; and finally, the evaluator agent evaluates whether the output of the forward reasoning thinking chain exceeds a medical conflict threshold, and iteratively updates an output processing conclusion. According to the method, complex problems in medical information processing problems can be analyzed, and reasoning accuracy and interpretability are remarkably improved.

The invention specifically relates to the technical field of natural language processing, and discloses an intelligent interactive customer service system based on an AI large model, which comprises a multi-modal input analysis module, an intention preliminary detection module, a large model semantic understanding module, a dynamic knowledge base association module, a dialogue strategy generation module, a natural language generation module and a service feedback work order module, through deep cooperation of a multi-modal input analysis module and a large model semantic understanding module, entity and attribute information is accurately extracted, context semantic association information is mined, and through the cooperative effect of an intention preliminary detection module, a dynamic knowledge base association module and a dialogue strategy generation module, based on a semantic analysis result output by a large model, a dialogue strategy is generated. According to the method and the system, matching calculation is carried out with the constructed dynamic knowledge base, rich knowledge support is provided for solving complex problems, personalized knowledge content is recommended for different users through the natural language generation module and the service feedback work order module, and the reply matching degree and the complex problem processing capacity are greatly improved.

The invention provides a cue word hierarchical optimization method for complex problem decomposition, which belongs to the field of artificial intelligence and natural language processing, and comprises task analysis, hierarchical cue word generation, feedback optimization mechanism, example reinforcement learning and multi-round interactive optimization. The execution effect and adaptability of complex tasks are improved, and the reasoning quality and the stability of generated results are improved.

The invention discloses a customer service robot system integrating knowledge base question and answer retrieval and work order processing, which comprises a natural language understanding and generating module, a knowledge retrieval module, a business system integration module, a dialogue management and decision module and a cue word management module, and relates to the technical field of artificial intelligence. Through cooperation of a large language model and a dialogue management module, the system realizes multiple rounds of deep understanding of a natural language, and can actively guide a user to clarify a demand, such as gradually inquiring an equipment model and fault details during repair, memorizing a dialogue context and avoiding information omission. Compared with traditional regular customer service, the method has the advantages that the accuracy of understanding complex problems is remarkably improved, open dialogue scenes such as technical consultation and fault diagnosis are supported, user interaction experience is closer to manual service seats, and the problems that multi-round dialogue understanding is limited, progress data cannot be obtained in real time, multiple knowledge bases are fused, and the automation degree of the business process is low are solved.

The invention belongs to the technical field of knowledge graph and big language model combination, discloses a big language model inference enhancement method based on structure perception knowledge graph representation, and provides a structure perception knowledge injection framework which utilizes a graph neural network to combine with multi-source structure information coding to enhance entity representation. And through a contrast learning mechanism, realizing alignment of semantics and a structure space, and meanwhile, designing an adaptive robustness fusion pruning training strategy, dynamically screening high-quality evidence sub-graphs, and guiding a language model to carry out structured reasoning. According to the method, deep integration and optimal utilization of the structure information and the semantic information are realized, the accuracy, robustness and resource efficiency of the model under a complex query task are remarkably improved, and a stronger embedding basis is provided for generation quality, so that the generation accuracy, stability and efficiency under a complex problem are improved.

The invention discloses a low-altitude atmosphere three-dimensional wind field inversion method based on wind profile radar vertical observation, relates to the technical field of three-dimensional wind field inversion, and aims to solve the problem of poor inversion effect caused by inaccurate model parameters in the wind field inversion process. The difference between simulation data and actual observation data is evaluated from multiple angles, a comprehensive optimization basis is provided, parameter calculation optimization is performed by adopting a particle swarm optimization method, the method has the characteristics of high convergence speed and high global search capability, is suitable for optimization of complex problems, and is used for confirming model requirements according to a wind field inversion rule. The model is selected by combining the evaluation result of the model library, it is ensured that the selected model has high pertinence and applicability, the prediction precision and reliability of the model are improved, model initialization is conducted through wind field fusion data, data information of multiple sources is fully fused, and the comprehensiveness and accuracy of the data are improved.

The invention relates to a multi-agent-based material performance prediction and synthesis system, and the system comprises a multi-agent data enhancement module which is configured to be used for firstly disassembling a complex problem into a plurality of subtasks, and then constructing a fine tuning data set comprising Sub-CoQ question and answer pairs by starting multi-source parallel retrieval; the multi-expert debate module is configured to be used for simulating decision conflicts of different roles in material engineering and generating a direct preference optimization DPO data set through debate; the training and verification module is configured to be used for training and verifying a large model MatMind in the field of materials by utilizing supervised fine tuning SFT and reinforcement learning RLHF based on the fine tuning data set and the DPO data set; and the material performance prediction and synthesis module is configured to be used for realizing intelligent recommendation of a material performance prediction and synthesis process by importing input parameters into the large model MatMind.

The invention discloses a knowledge-based visual question-answering method and device based on big and small model collaboration and a medium, belongs to the technical field of visual question-answering, and solves the problem of how to improve the accuracy of complex questions faced by a visual question-answering technology. In the image description extraction step, an image description and key object label set which is highly associated with a natural language problem is generated, and in the entity enhancement processing step, key entities are extracted and a clarity problem and an entity example which are used for resolving semantic ambiguity are generated. In the candidate answer generation step, a plurality of semantically complementary answers are generated under different reasoning dimensions on the basis of a multi-round generation strategy, optimized candidate answers are screened on the basis of a model built-in scoring mechanism, in the context example retrieval step, examples most related to current input are retrieved, and finally, the candidate answers are obtained. And the large language model receives a unified input prompt formed by splicing output results in the steps, and inference is performed in an autoregression mode to generate answers, so that the accuracy of the visual question-answering technology facing complex questions is effectively improved.

The invention discloses a dynamic retrieval enhanced generation and deep reasoning method and framework based on an intelligent agent, and belongs to the technical field of natural language processing, and the method comprises the following steps: S1, inputting a problem into a big language model which analyzes the problem and divides the problem into three difficulty levels, the problems comprise a simple difficulty problem, a medium difficulty problem and a complex problem; s2, respectively processing the problems according to different difficulties of the problems, and retrieving related knowledge from a corpus; and S3, processing the knowledge obtained in the S2 by the large language model, and outputting a final result. According to the dynamic retrieval enhanced generation and deep reasoning method and framework based on the intelligent agent, the performance of a large language model in various complex tasks is enhanced, especially in the aspect of multi-hop reasoning, the overall efficiency and accuracy of a system are improved through the synergistic effect between modules, and the system is suitable for large-scale popularization and application. And the method is suitable for various knowledge-intensive application scenes.

The invention discloses an AI (artificial intelligence) agent construction system based on hybrid retrieval and father-child segmentation, which comprises the following steps of: dividing a subclass knowledge base according to domain knowledge, performing father-child segmentation processing, and constructing a hierarchical semantic network; vectorization embedding and deep semantic reconstruction are carried out on the user question text; retrieving the reconstructed problem by adopting a mixed retrieval algorithm combining sparse retrieval and dense retrieval, and forming a high-score sub-segment set according to a comprehensive score obtained by dynamic weight distribution; mapping the sub-segments to the parent segment through a hierarchical backtracking algorithm, aggregating brother nodes to form an extended candidate set, and generating an associated sub-segment set after duplicate removal and re-retrieval; and finally inputting a large language model to generate a complete answer. According to the method, the problems of context segmentation, low retrieval accuracy and complicated knowledge base maintenance of traditional document segments are solved, the answer coverage and accuracy of an intelligent question-answering system are remarkably improved, and the method is suitable for knowledge question-answering scenes in the complicated technical fields such as intelligent network connection automobiles and the like.

The invention relates to the technical field of reinforcement learning, in particular to a language model reinforcement learning training extension method and device based on reflection, and the method comprises the steps: inputting a target natural language reasoning problem into a pre-constructed large language model, so as to generate a target reflection thinking chain of the large language model; based on the target reflection thinking chain, performing reinforcement learning training expansion on the large language model to obtain a target inference model; and evaluating the relationship between the target performance and the generation length of the target inference model to obtain an evaluation result, and determining the target effectiveness of the inference expansion of the target inference model by using the evaluation result. Therefore, the problems that the practical understanding of large-scale reinforcement learning training is limited and the ability of a large language model to solve complex problems is reduced due to less extension potential and attribute attention of reinforcement learning from human feedback in related technologies are solved.

The invention discloses a network management racing item question answering system based on a knowledge graph, and belongs to the technical field of knowledge graph teaching. In the system, the association relationship among massive network equipment, protocol configuration and fault cases is stored in a knowledge graph structuralization mode, and natural language processing and graph database retrieval technologies are combined; core knowledge points of complex problems can be quickly positioned, and strong scene adaptability and expansion potential are shown through modular design and a dynamic cooperation mechanism. The knowledge graph management module supports incremental updating and cleaning fusion of multi-source heterogeneous data, can integrate newest competition cases and technical specifications in time, and guarantees the frontier and authority of a knowledge base; and the interaction optimization module automatically identifies a knowledge blind area according to the score of the user on the answer and triggers graph completion. The self-feedback and self-iteration mechanism not only reduces the manual maintenance cost, but also provides a technical basis for cross-domain knowledge migration.

The invention relates to a double-layer double-objective optimization scheduling method based on a large language model, and the method comprises the steps: building a double-layer double-objective optimization model in a server, and building an upper-layer model and a lower-layer model for an AISTS problem based on a plurality of different variables, different constraint conditions and different objective functions; and prompting a dual-objective optimization model of adjustment and algorithm evolution to form a two-stage method based on LLM, the method can overcome the defects that a traditional algorithm needs to design the algorithm manually and adjust parameters manually, the calculation cost is high, the efficiency is low, complex problems are difficult to effectively decompose and coordinate double-layer optimization, a dynamic collaborative optimization framework is lacked, and the flexibility of algorithm combination is low.

The invention discloses a method for sharing a host based on multiple input devices, and belongs to the technical field of computers, and the method comprises the following steps: after a management user logs in a multi-seat configuration tool, creating second seats with the same number as new input devices according to a Udev rule, and enabling the management user to be a user operating the input device corresponding to a first seat of the host; according to a binding instruction sent by the management user, the second seats and the new input devices are bound one by one, and operation users of the input devices corresponding to the second seats are common users; according to a permission configuration instruction sent by the management user, configuring an application window permission for each common user; and sending the binding relationship between the second seat and the new input device and the application window authority to a window manager, so that the window manager manages the application window according to the binding relationship and the application window authority. According to the invention, the problems of disordered management and complex configuration when multiple users share one host window are solved, and more flexible and efficient equipment management experience is provided for the users.

A multi-objective optimization method for a distribution network containing distributed power sources taking into account carbon emissions, comprising the following steps: 1) constructing a multi-objective optimization model for the distribution network; 2) using a hierarchical analysis method to determine the weights of each objective function in the above-mentioned multi-objective optimization model; 3) using an improved beetle whisker particle swarm algorithm to improve the global search capability and local search accuracy of the algorithm for solving the above-mentioned multi-objective optimization model. By integrating a multi-objective optimization model and an improved hierarchical analysis method, the present invention not only optimizes economic indicators, but also takes carbon emission costs into consideration for the first time, thereby achieving the dual economic and environmental benefits of distribution network planning. The improved beetle whisker particle swarm algorithm used further improves the search efficiency and solution accuracy, ensuring the robustness and rapid convergence of the algorithm when dealing with high-dimensional complex problems. The method effectively reduces network losses and carbon emissions, improves voltage distribution, and provides technical support for the green transformation and sustainable development of power systems.

The invention discloses a large model training method, a question and answer method, related equipment and a computer program product. The method comprises the steps of obtaining question and answer training data; the problem samples are sent to a to-be-trained large model for reasoning, and prediction output of the large model is obtained; calculating an exploration reward based on hidden layer state characteristics generated in a big model reasoning process, wherein the exploration reward is used for encouraging the big model to process the problem sample by adopting an unknown reasoning path; calculating a result reward based on the predicted output and the answer tag; and updating parameters of the large model by adopting a reinforcement learning mode according to the exploration rewards and the result rewards. Exploration rewards are additionally added in the reinforcement learning process, a large model can be encouraged to explore an unknown reasoning path, local optimum is avoided, the probability that the large model discovers a better reasoning path when facing a complex problem is improved, the performance of the large model on a complex reasoning task is improved, and the reasoning efficiency of the large model is improved. And the accuracy of the answering result of the complex reasoning question is improved.

A reference point is arranged on the ground to calibrate a ground world coordinate system, a robot base coordinate system is established based on robot base features, and a rigid transformation matrix between the two is calculated to achieve spatial alignment. A multi-sensor network is constructed, a visual sensor and a laser radar are associated to a double-reference coordinate system, and data are synchronously collected and unified to the same coordinate system. And according to the spatial consistency error of the data, dynamically evaluating the credibility weight of the sensor, constructing a joint error equation, and based on least square method optimization, weighting fusion data and outputting calibration parameters. According to the invention, the problems of dependence on single sensor precision, lack of a system-level calibration scheme and complex calibration process are solved, and the calibration precision and efficiency of the robot system are improved.

The invention discloses a use method of a physical information neural network for an optimal control problem, and belongs to the field of optimal control problems. The method comprises the following steps: adding physical constraints into a loss function of a physical information neural network, and dynamically adjusting weights according to loss changes in a training process through a self-adaptive weight mechanism, so that the network pays more attention to an area with relatively large loss, thereby improving the solving precision of a model for complex problems. According to the method, under the condition of not needing any data, by fusing constraints of a physical equation and using a self-adaptive weight mechanism, the optimal control law is more accurately approached, a traditional method for refining a grid is replaced, and the problem that in the face of two-dimensional, three-dimensional or high-dimensional problems, the optimal control law is more accurately approached by the traditional method is solved. The calculation amount is large due to the fact that the grids need to be refined continuously.

The invention discloses a large model-knowledge graph driven data blood relationship traceability reasoning method and system. According to the method, through a closed-loop framework of KG-driven dynamic decomposition, small-model lightweight reasoning and visual credibility verification, the defects are systematically solved, and an interpretable, low-cost and easy-to-expand data consanguinity depth traceability scheme is provided for scenes with high credibility requirements such as data governance, compliance auditing and root cause analysis. According to the system, aiming at complex problem requirements, after a complex problem is input, the problems of no interpretation, low reliability and the like in data blood relationship traceability are solved through two modules, namely problem reasoning and verification presentation. The modular design not only conforms to the technical decoupling principle, but also facilitates independent optimization, and is suitable for medical treatment, scientific research and other scenes with high requirements for interpretability.

The invention relates to a digital textbook knowledge explanation method based on a time sequence knowledge graph, which comprises the following steps: receiving an explanation request from a user side, and associating the explanation request with a specific teaching content unit in a pre-constructed time sequence knowledge graph by performing intention analysis on the explanation request; based on the current learning progress of the user and a target teaching content unit corresponding to the explanation request, extracting a knowledge explanation constraint sub-graph conforming to the sequential logic from the sequential knowledge graph; and based on the knowledge explanation constraint subgraph, generating explanation content conforming to the sequential logic, and outputting the explanation content to a user side. The explanation content is strictly constrained to the current learning progress of the student, concept outlines are avoided, and the explanation accuracy and acceptability are improved; a complex problem solving method is systematically decomposed and is clearly explained according to a teaching sequence, so that the understanding quality is improved; the confusion of students caused by outline exceeding or complex steps is reduced, and the learning efficiency and effect are improved.

The embodiment of the invention discloses a 4K splicing processor and a 4K splicing display system. The 4K splicing processor comprises a driving circuit used for processing a 4K resolution signal, the driving circuit comprises an FPGA chip U5, TVS array chips U10, U11, U12 and U13 and DP signal line connectors CN9 and CN10; the model of the U5 is Spartan-6, the models of the U10, the U11, the U12 and the U13 are all RClamp0524P, the CN9 is connected with the U5 through the U10 and the U11, and the CN10 is connected with the U5 through the U12 and the U13; the U5 receives the input 4K resolution signal through the CN9 to process the 4K resolution signal and outputs the processed 4K resolution signal through the CN10 to achieve 4K resolution signal looping. According to the 4K splicing processor and the 4K splicing display system, the problem that 4K splicing display system equipment is complex is solved.

The invention discloses a query statement splitting method and device, equipment, a medium and a product. The method comprises the steps of obtaining a to-be-queried text, a pre-constructed prompt project and a knowledge graph based on a graph database; inputting the to-be-queried text into the large language model; and splitting the to-be-queried text into natural sub-question sets in a primitive interpretation mode based on the prompt project and the knowledge graph through the large language model to obtain an output result of the large language model. The method comprises the following steps: defining a primitive pre-construction prompt project, guiding a large oracle model to convert a primitive into a natural language for explanation, and disassembling a complex problem corresponding to a to-be-queried text into sub-problems in a primitive explanation mode; the problem of difficulty in understanding primitives of a large language model is solved, and accurate understanding of user query intention and semantics is realized, so that the accuracy of splitting the to-be-queried text question is ensured, and a basis is provided for subsequent generation of accurate query answers.

The invention provides an automatic data analysis method based on multi-agent cognitive collaboration, which comprises the following steps: S1, multiple agents perform modeling processing, and each agent has unique functions and tasks in a multi-agent system; s2, the user side is used for collecting complex problems proposed by a user; s3, the controller coordinates multiple agents to work, wherein the multiple agents comprise a problem analysis module, an SQL module, an OA module and a data analysis module; s4, analyzing a complex problem into a plurality of sub-problems by a problem analysis module; s5, the SQL module searches a table structure according to the sub-problem requirements, compiles an SQL and operates the SQL; s6, the QA module is responsible for checking whether the sql is correct or not; according to the method, the search technology and the large model generation technology are combined, knowledge outside the model can be obtained through the search technology, a tool can be directly operated to solve user problems, and more complex problems can be solved in multiple rounds of interaction.

The invention discloses an unmanned aerial vehicle cluster safety path planning method based on an improved black-wing optimization algorithm, and the method comprises the steps: introducing an anti-collision constraint for each unmanned aerial vehicle in an unmanned aerial vehicle cluster, and forming a multi-factor constraint flight total cost function of the unmanned aerial vehicle cluster; introducing a good point set theory, an optimal value guide differential mutation strategy and an elite learning mechanism to improve a black wing plinary optimization algorithm; taking a multi-factor constraint flight total cost function of the unmanned aerial vehicle cluster as a fitness function, and solving the fitness function by adopting an improved black wing plinth algorithm; and simulating the path planning of the unmanned aerial vehicle, and verifying the effectiveness of the proposed algorithm. According to the algorithm, better performance can be obtained in the face of complex problems; the algorithm has flexible adjustment capability, the path is updated according to the information obtained in real time, learning can be carried out by comparing the difference between the algorithm and elite individuals and the difference between the algorithm and the population average value in the path planning process, and the algorithm performance is improved.

The invention discloses a multi-agent-based context knowledge internalization, sharing and multiplexing system and method. The system comprises a multi-source knowledge capture module, a situation information extraction and structuring module, a knowledge internalization and incremental learning module, a situation awareness knowledge reuse and recommendation module and a knowledge quality evaluation and maintenance module, wherein each module is composed of a special intelligent agent or a functional unit and works cooperatively; according to the method, interaction, cooperation and process data and expert input of a task execution full cycle of a multi-agent system are captured through a multi-source module, preprocessing is carried out through a situation extraction module, a situation is associated to generate structured knowledge, and then the structured knowledge is fused into a dynamic knowledge system through an internalization module to achieve incremental learning. In a new task, matching knowledge is retrieved and recommended through the context awareness module, and finally the knowledge quality is periodically maintained by the evaluation module to ensure accuracy and timeliness; according to the method, the autonomous learning ability, the knowledge reuse efficiency and the intelligent level of solving complex problems of the multi-agent system are remarkably improved.

The invention provides a knowledge question and answer model construction method. The method comprises the following steps: taking a pre-training language model as a base model, and executing two-stage training according to the following modes to obtain a knowledge question and answer model: function understanding pre-training: constructing a plurality of function definition codes and a plurality of function call examples to train the base model so as to enable the model to learn a function execution principle; and instruction fine tuning training: obtaining a task instruction, a query instance and a query program corresponding to the query instance to train the initial model so that the model learns a function combinatorial programming principle. According to the technical scheme, the knowledge question and answer model is trained by proposing the two-stage training strategy, so that the model has the function understanding capability for knowledge programming and the capability of converting a complex problem into a query program based on the function understanding capability, and the problems that in the prior art, the model tends to memorize the program, and the program is difficult to query are solved. And therefore, the generalization of the model for the unseen query semantics contained outside the domain is improved.

The invention discloses a deep learning-based human behavior recognition method and recognition system, and the method is characterized in that the method comprises the following steps: S1, data collection: collecting video data containing a plurality of human behaviors through a high-definition camera or a video database, and covering different scenes, illumination conditions and human postures; s2, data preprocessing: firstly extracting a three-dimensional coordinate sequence of human skeleton key points by using a Medape pre-training model, and then performing mean filtering and normalization preprocessing on the collected data to improve the generalization ability of the model; according to the method, the advantages of the DCN and the LSTM are combined, the improved double-flow convolution model is added with the space-time soft attention mechanism, the problems that space-time feature extraction is insufficient and feature design is complex in human body action recognition are effectively solved, the lightweight degree of the model is remarkably improved, and the human body action recognition accuracy is improved. And a new method is provided for research and application of human behavior recognition.