12 results about "Rule based inference" patented technology

The invention relates to a cloud computing-based enterprise digital tool chain dynamic collaboration and deployment method, which comprises the following steps of: according to a data mapping relationship change list, analyzing change contents of an interface protocol and a data field by adopting a rule-based reasoning method, and according to a pre-configured interface protocol adaptation rule, establishing a data field according to the change contents of the interface protocol and the data field; generating data mapping scripts among different versions of tools; after the discarded interface is deleted, a graph-based shortest path search algorithm is adopted, the deleted discarded interface is bypassed, a data transmission path between tools is re-planned, and a complete tool chain data transmission path is re-formed; and adjusting the size of the data fragments and transmission batch parameters, optimizing the circulation efficiency of the data in the tool chain, and dynamically adjusting the execution sequence and the concurrency of each tool in the tool chain according to the optimized data circulation scheme.

The invention discloses a business logic reduction method and system based on spatio-temporal data, and belongs to the field of spatio-temporal data, and the method comprises the steps: constructing a feature sample library and a field-based formalized logic reasoning engine, employing a rule-based reasoning method, achieving the cooperative work among a plurality of reduction, and forming a complex and flexible business logic; each protocol is regarded as a spatio-temporal data cube, and a protocol collaboration mechanism based on data feature gridding association is introduced by designing a reward function, so that the protocol collaboration mechanism learns and optimizes own protocol strategies and balances local optimization and global optimization targets in the collaboration process with other protocols; the optimized protocol strategy is applied to the protocol management method based on the micro-service architecture, different protocol rules are packaged into independent micro-services, communication and cooperation between the services are achieved through a standardized interface, and the protocol management method adapts to evolution and change of service logic.

The invention discloses a feature and attribute structured compilation method and system based on multiple kinds of real estate voucher data, and belongs to the field of feature and attribute structured compilation, and the method comprises the steps: obtaining various kinds of voucher data in the real estate field, extracting the key information of a voucher for each kind of voucher, and constructing a voucher information feature vector; for a plurality of credible certificates, key information in the certificates is extracted, structured analysis is carried out on the certificate information through a natural language processing technology, and a unified real estate information model is constructed; in order to solve the problems of information conflict and inconsistency between different vouchers, a rule-based reasoning method is adopted, real estate domain knowledge is combined, conflict solving rules are set, conflict information is verified and corrected, and information consistency is guaranteed.

The application relates to a fault diagnosis method of an aviation equipment system based on rule reasoning and a knowledge base, which comprises the following steps: S1, constructing a fault diagnosis data model of the aviation equipment system, acquiring fault diagnosis information, and establishing a fault knowledge base; S2, determining associated fault diagnosis information of the aviation equipment system based on rule reasoning of a rule base, and acquiring a similarity result through the knowledge base; S3, analyzing the similarity result of the fault diagnosis information of the aviation equipment system, sorting, and outputting a fault diagnosis result. The application extracts and identifies fault features based on a keyword base; obtains matched associated LRUs based on rule reasoning, and records corresponding trace information; searches and matches fault diagnosis information based on the knowledge base, calculates information matching similarity, completes similarity sorting of a final fault result, and determines the fault diagnosis result of the aviation equipment system. The application improves the fault diagnosis efficiency of the aviation equipment system, reduces the fault diagnosis time, and prolongs the service cycle of the aviation equipment.

The invention discloses a diffusion model efficient fine tuning method for open recognition, and relates to reasoning space compression based on a Top-K rule. Constructing a local sorting graph based on HodgeRank and accelerating posteriori estimation; generating an open sample based on semantic consistency; on the basis of joint discriminant optimization of consistency indexes, on the premise that closed category classification accuracy is kept, the sampling number and calculation time needed by the model in the reasoning stage are remarkably reduced, and reasoning efficiency is greatly improved; and meanwhile, through explicit modeling, a semantic boundary and a discrimination mechanism of the samples are opened, so that the recognition and rejection capabilities of the model on the samples with unseen categories are effectively enhanced. In conclusion, the method realizes good balance among accuracy, efficiency and open generalization ability, and has wide application value and popularization potential in industrial deployment, scientific research application and high-risk identification tasks.

The invention provides a dental implant osseointegration prediction method and system based on dynamic data, and the method comprises the steps: carrying out the feature extraction of preoperative risk assessment data through a time sequence analysis method according to an initial data set, and obtaining a preoperative osseointegration risk feature set; according to the synostosis state change trend, clinical decision support information is generated by adopting a rule-based reasoning method, and the treatment priority of the current stage is determined; if the treatment priority is higher than a preset threshold value, extracting prediction information related to the priority from the dynamic data set through a real-time information matching algorithm to support decision making of doctors; through a real-time matching result of prediction information, a decision tree algorithm is adopted to update supply rule design, and optimized clinical decision support information is obtained; and according to the optimized clinical decision support information, continuously updating postoperative healing monitoring data by adopting a feedback circulation mechanism to obtain a latest osseointegration quality evaluation result.

The application discloses a kind of based on multi-agent reinforcement learning's car body clamp clamping unit generation method. Among them, the method includes: generating the standardization input data of car body clamp clamping unit, wherein the standardization input data includes RPS grouping prediction result, RPS distance margin parameter and car body part structure feature;Based on the standardization input data, the standardization input data is inferred by graph neural network or rule-based reasoning engine, and parameter constraint subgraph containing part node, constraint relationship and assembly logic is generated;Based on the parameter constraint subgraph, the parameters of the car body clamp clamping unit are optimized by multi-agent reinforcement learning model, and the optimization parameter vector is obtained, and the car body clamp clamping unit is generated based on the optimization parameter vector.The application solves the technical problems that the design of existing clamp clamping unit is based on static library, lacks dynamic relationship inference and cannot adaptively assemble.

The present application relates to a cloud computing-based enterprise digital tool chain dynamic collaboration and deployment method. According to the data mapping relationship change list, a rule-based reasoning method is used to analyze the changes in interface protocols and data fields, and according to the pre-configured interface protocol adaptation rules, a data mapping script between tools of different versions is generated; after deleting the obsolete interface, a graph-based shortest path search algorithm is used to bypass the deleted obsolete interface, re-plan the data transfer path between tools, and re-form a complete tool chain data transfer path; adjust the data segment size and transmission batch parameters to optimize the data flow efficiency in the tool chain, and dynamically adjust the execution order and concurrency of each tool in the tool chain according to the optimized data flow plan.

The invention discloses a vehicle body clamp clamping unit generation method based on multi-agent reinforcement learning. The method comprises the steps that standardized input data of a vehicle body clamp clamping unit is generated, and the standardized input data comprises an RPS grouping prediction result, an RPS distance edge distance parameter and vehicle body part structure characteristics; based on the standardized input data, reasoning the standardized input data through a graph neural network or a rule-based reasoning engine, and generating a parameter constraint sub-graph comprising part nodes, constraint relationships and assembly logic; and on the basis of the parameter constraint subgraph, parameters of the vehicle body clamp clamping unit are optimized through a multi-agent reinforcement learning model, an optimized parameter vector is obtained, and the vehicle body clamp clamping unit is generated on the basis of the optimized parameter vector. The technical problems that the design of an existing clamp clamping unit is established on the basis of a static library, dynamic relation reasoning is lacked, and self-adaptive assembly cannot be achieved are solved.

The invention relates to the technical field of health monitoring, in particular to a big data-driven patient health monitoring method, which comprises the following steps of: acquiring health data of a patient through medical equipment and an electronic medical record system, including conventional health parameters and rare health parameters; based on an improved self-organizing mapping algorithm, introducing a neighborhood dynamic adjustment mechanism to update a neighborhood function, and performing pattern mining on the health data; in combination with a known health data template and an unsupervised data pattern mining result, generating annotation data through a rule-based reasoning and fuzzy matching technology; and according to the labeled data, combining historical health data of the patient, and utilizing a big data analysis prediction model to generate a personalized health risk assessment report. According to the method, high-sensitivity recognition of rare health parameters is realized, and the defects of a traditional method in rare data recognition are overcome.

The invention provides a spacecraft overall design large model training method based on a knowledge graph, and relates to the field of spaceflight. The problems that in the aerospace field, effective data in model training is insufficient, data generated by multidisciplinary collaborative design has heterogeneity, a traditional expert system is limited by rule-based reasoning and a preset interpreter, and design exceeding a preset rule and cross-subsystem is difficult to solve are solved. The method comprises the steps of obtaining a spacecraft overall design knowledge graph; a deep entity disambiguation model based on BERT is constructed; and training a deep entity disambiguation model based on BERT through an SFT algorithm. The method is applied to the field of spacecraft design.

An edge-optimized retail analytics system for autonomous in-store decision support, consisting of: an edge computing unit housed in a robust, thermally conductive enclosure and configured for use in a retail store environment; System-on-Module (SoM) mounted on a multilayer printed circuit board (PCB), wherein the SoM comprises a multi-core central processing unit (CPU) configured to manage sensor data orchestration and rule-based inference processing, and a neural processing unit (NPU) configured to perform deep neural network inference operations in real time; a high-bandwidth volatile memory module electrically connected to the SoM to buffer time-aligned multimodal sensor streams; a non-volatile solid-state drive (SSD) configured to persistently store AI model weights, inference outputs, and business-specific event logs; a power management circuit integrated on the printed circuit board, configured to regulate the input voltage of a Power-over-Ethernet (PoE) line; and a sensor interface bus that is connected to a variety of sensor modules, including visible spectrum cameras, thermal imaging sensors, passive infrared motion sensors, RFID readers, and load cell-based shelf weight sensors.