216 results about "Feedback loop" patented technology

An AI-powered content generation system that creates consistent, coherent, and engaging multi-modal content by integrating multiple specialized AI components. The system analyzes user input, identifies key elements, and maintains continuity throughout the generation process. It incorporates a feedback loop to learn and adapt based on user preferences, enabling personalized content experiences. The modular architecture allows for seamless integration of AI components focusing on text, images, audio, and interactive elements. The system ensures consistency across modalities and over extended periods, while managing rights, licenses, and royalties using blockchain technology. This advanced platform revolutionizes content creation, consumption, and management in the digital age.

The invention relates to the technical field of communication, in particular to a dynamic priority dual-mode communication fault-tolerant switching method and equipment, and aims to solve the problems of communication interruption and service quality reduction when a main link is degraded in a complex electromagnetic environment. According to the method, a dynamic priority evaluation model and a dual-mode protocol cooperation mechanism are constructed, throughput, bit error rate, delay jitter, task semantic priority and aging constraint data of a main link and a standby link are collected in real time, a priority weight and a communication mode decision matrix are generated through combined modeling, an atomization switching instruction is executed according to the priority weight and the communication mode decision matrix, and a communication mode is switched. And completing path reconstruction and resource redistribution. The system further introduces link health prediction, a resource reservation pool and an online learning feedback loop, and supports fault pre-judgment, key task preemption and model adaptive optimization. According to the method, the limitation of a static threshold is broken through, the task semantics and channel state joint scheduling is realized, and the communication reliability, the service continuity and the intelligent fault-tolerant capability of a distributed system in a high-risk scene are remarkably improved.

The invention discloses a structured data quality evaluation and optimization system using a large language model, which comprises a data access module used for formatting and preprocessing data, a data quality detection module used for establishing data quality evaluation indexes, performing statistical analysis and rule verification to detect the data and calculating the overall quality grade, and a big language model, the semantic analysis and SQL generation module is used for analyzing a database structure through a large language model, converting a natural language instruction into an executable SQL statement and is based on the large language model, and the anomaly detection and optimization suggestion module is used for positioning anomaly by utilizing a mixed detection algorithm and generating natural language interpretation and optimization suggestions in combination with the large language model. And the user feedback and dynamic adjustment module is used for absorbing adjustment suggestions through a feedback loop mechanism, and optimizing a prompt project and an SQL generation strategy. According to the method, semantic analysis is carried out on a database structure by adopting a large language model, and high-precision SQL query is automatically generated.

A system and method for adaptive data processing combining compression and encryption. The system analyzes input data characteristics, compares probability distributions, and creates a transformation matrix to convert data into a dyadic distribution. It generates a main data stream of transformed data and a secondary stream of transformation information. The system dynamically selects and applies processing techniques, including transformation, encoding, compression, and encryption algorithms, based on analyzed characteristics and real-time performance metrics. It compresses the main data stream using Huffman coding and implements security measures to protect the output. A feedback loop monitors technique effectiveness, updates a knowledge base, and influences future selections. The system can operate in lossless, lossy, or modified lossless modes, adapting to different application requirements. This approach offers an efficient solution for scenarios where both data reduction and security are critical concerns.

The present invention discloses an artificial intelligence-based (AI-based) system and method for generating optimised operation planning and scheduling output. The AI-based system obtains at least one of: one or more data explanation videos, one or more process understanding videos, and unconstrained operational planning data, along with one or more prompts as an input. The AI-based system extracts one or more informative image frames and audio data, to train the one or more AI models and generate a planning standard operating procedure (SOP). The AI-based system processes the planning SOP, the constrained operational planning data, and the one or more prompts to generate the optimised operation planning and scheduling output based on an optimised function with a continuous feedback loop in response to at least one of: the one or more prompts, updated planning SOP, and real-time changes in the constrained operational planning data.

The invention relates to the technical field of audio and video emergency transmission, and discloses an audio and video low-delay return method and system in an extreme environment. The method comprises the following steps: acquiring original multi-modal data of audio and video acquisition equipment in a target area, and analyzing a data state of the original multi-modal data; and meanwhile, available network transmission links are monitored, and the quality is evaluated. Self-adaptive coding parameters are generated in combination with the link quality and the data state, dynamic coding is executed on original data, and a coding stream suitable for redundant transmission is formed. And carrying out cooperative distribution transmission based on the network link set and the coded stream to generate return data. According to the generation process, a device and energy management instruction is formed and issued to the acquisition and relay device. According to the method, the network state and the content characteristics are optimized in a dynamic coding link in a collaborative manner, and the audio and video quality, the real-time performance and the system energy efficiency which are transmitted back in an extreme environment are improved through a feedback closed loop for transmitting a result to equipment management.

The invention provides an outbound routing method and system based on multi-dimensional index coupling and large model decision, and the method comprises the steps: carrying out the preprocessing of the historical data of each line provider, predicting the future performance index through an LSTM neural network, and optimizing the prediction result through a Kalman filter, and eliminating the noise. In the route decision-making link, the system calculates the Q value of each line provider based on the deep Q learning network, comprehensively considers the call completing rate, average cost, negative feedback rate and other factors of each line provider, and intelligently selects the optimal line provider in combination with the user portrait, scene recognition and real-time load state. The system also constructs a complete feedback closed loop, continuously optimizes the model by collecting call result data, and automatically adjusts decision parameters. The call completing rate and the service quality of the outbound service are improved, and the communication cost is reduced.

The embodiment of the invention discloses a method for intelligently assisting in reviewing a water conservation scheme of a water and soil conservation production and construction project, and the method achieves the intelligent feature matching of a new scheme and similar cases through the construction of a tagged knowledge base integrating historical cases of multiple industries, so as to recognize the generality compliance. A double-engine review mechanism is adopted, a form review module automatically checks document structure integrity and format specifications, and a logic review module executes cross-chapter data consistency check and business logic compliance analysis and highlights and positions error positions in a document. And recording decision logic in real time to form a feedback closed loop. And dynamically optimizing case library matching weight and recognition rules based on high-frequency error statistics and expert correction records. The modes of positioning review key points to chapters and sections, describing specific requirements of each review key point in a language, adding a check formula among data to review content needing to be calculated, determining a check type to which each review key point belongs, and the like are adopted, and the model is trained more clearly, so that the model has relatively professional review ability.

The invention relates to the technical field of information processing, and discloses a method for dynamically regulating and controlling production parameters of a plastic film by utilizing program process scheduling. The method comprises the following steps: converging a real-time parameter sequence to a digital twinborn body to establish virtual mapping, and constructing a multi-dimensional data model; analyzing a material state evolution trend in the digital twinborn body, synchronizing with a physical production line state, fusing historical data to predict a potential influence range, and generating a regulation and control demand report; generating an adjustment scheme draft, performing priority ranking on simulation tasks in the adjustment scheme draft by adopting a program process scheduling algorithm, dynamically allocating computing resources, executing the simulation tasks in parallel, and determining an optimization parameter set in combination with a gradient descent algorithm; and iteratively correcting the parameters in the digital twinborn body through feedback circulation to generate a final coordination control instruction, inputting the final coordination control instruction into the digital twinborn body, confirming synchronism and outputting a report. According to the invention, the precision of dynamic regulation and control of the production parameters of the plastic film is improved.

The invention discloses a medical visual question and answer model construction method based on a feedback loop. The method comprises the following steps: step 1, generating initial candidates; step 2, performing multi-dimensional candidate evaluation; 3, iterative optimization of feedback driving is carried out; step 4; and 5, efficient fine adjustment of model parameters: performing fine adjustment on a pre-trained multi-modal large language model by using the enhanced training data set constructed in the step 4 to obtain a final medical visual question and answer model. According to the method, a feedback loop mechanism is introduced, and cooperative and iterative self-correction and optimization are carried out on generated text interpretation (Ratio) and a visual localization region of interest (RoI), so that the fact reliability, the logic consistency and the visual strong relevance of a final output result are systematically ensured.

A computer-implemented system and method for dynamic, explainable, and adaptive prioritization of cybersecurity controls is disclosed. The system ingests unstructured threat reports and employs a natural language processing (NLP) module to automatically extract adversary tactics, techniques, and procedures (TTPs). A scoring module applies a mathematical time-decay function to the extracted intelligence. A novel hybrid prioritization engine provides explainability-by-design by computationally integrating these objective, data-driven scores with organization-specific context within a transparent multi-criteria decision analysis (MCDA) model. Critically, the system establishes a self-optimizing closed feedback loop; it receives real-world control effectiveness metrics from the operational environment and uses this data as new ground-truth labels to continuously and automatically retrain internal machine learning models. This adaptive mechanism improves the computer's own predictive accuracy and resource allocation efficiency over time, representing a tangible technical improvement.

According to the financial risk control risk assessment method based on the dynamic list library, the problems that a traditional method is poor in real-time performance, low in precision and insufficient in adaptability are solved through dynamic weight self-calibration list library construction, multi-model fusion of a self-adaptive scene, real-time assessment of dual-channel cross validation and a closed-loop self-adaptive response mechanism. The core innovation of the method lies in that a list library and model evaluation are deeply coupled, risk attenuation factors, dynamic weight adjustment and feedback closed loops are introduced, the fraud recognition rate is improved, the misjudgment rate is reduced, and the method can be widely applied to financial scenes such as banks and payment platforms.

The embodiment of the invention provides a big language model reasoning optimization system, method and equipment and a storage medium. The system comprises an asynchronous KV cache pre-fetching module, a heterogeneous Specific decoding module, a dynamic model re-fragmentation scheduling module, a Memory-Aware dynamic batch module and a scheduling control engine, and can organically integrate the asynchronous KV cache pre-fetching module, the heterogeneous Specific decoding module, the dynamic model re-fragmentation scheduling module and the Memory-Aware dynamic batch module, so that the dynamic model re-fragmentation scheduling module and the Memory-Aware dynamic batch module are integrated, and the dynamic model re-fragmentation scheduling module and the Memory-Aware dynamic batch module are integrated. And a set of closed-loop adaptive optimization system is formed under the unified coordination of a scheduling control engine, so that module-level linkage, adaptive coordination and performance feedback closed loop in the reasoning process are realized, and finally, the throughput, the response speed and the resource utilization rate in the reasoning stage are remarkably improved on the premise of not influencing the accuracy of a large language model.

The invention discloses an intelligent main plan scheduling system oriented to the chemical industry, and the system comprises a data center layer which is used for aggregating order, equipment, materials, inventory and process constraint data; the MPS optimizer operates the tabu search module, the step-counting hill-climbing module and the flood algorithm module in sequence or in parallel in the same optimization period, and automatically switches algorithms under the following conditions: when the continuous iteration improvement rate is lower than a set threshold value, the tabu search is switched to step-counting hill-climbing, and the step-counting hill-climbing is switched to step-counting hill-climbing; when the local search step number is not improved, starting the flood algorithm to re-initialize the candidate solution; and the process scheduler is used for splitting the coarse-grained plan output by the MPS optimizer into minute-level process schedules, and transmitting the execution deviation back to the data center layer in real time through a bidirectional feedback loop so as to trigger incremental rearrangement. According to the intelligent main plan scheduling system for the chemical industry, through a hybrid algorithm dynamic switching mode, the thousand-level order rearrangement time is greatly shortened, and the plan response speed is remarkably improved.

The invention discloses an incentive differentiation configuration and promotion support system based on user hierarchies, and relates to the technical field of big data and intelligent promotion, and the system comprises a multi-dimensional user data collection module which collects various types of user data and carries out desensitization storage; the user level dynamic evaluation module is used for calculating a comprehensive score through multi-dimensional index weighting and dividing five layers of users; the incentive differentiation configuration module formulates gradient incentive strategies according to different hierarchies; the promotion resource intelligent distribution module distributes promotion resources according to levels, targets and scenes; and the promotion effect tracking feedback module is used for monitoring data, generating a report and dynamically adjusting a strategy to realize optimization. According to the method, accurate division of user levels is realized, the user participation degree is improved through differential excitation, the utilization efficiency is improved through intelligent distribution and promotion of resources, continuous optimization is performed through an effect feedback closed loop, different business targets are adapted, and the marketing promotion accuracy and efficiency are improved.

The invention discloses a servitization arrangement and scheduling method and system based on a large language model. The method comprises the steps that a model and a working model of the model are packaged into an MSU, and a universal interface specification file is generated based on an MCP protocol; a unique Context ID is generated for each reasoning task, and context information is registered in an external shared storage; automatically generating a model service arrangement diagram by the LLM based on the UIS file of each MSU; according to the topological relation and context dependence of the arrangement diagram, the user-defined scheduler comprehensively scores candidate nodes according to context affinity and resource scores in a K8s cluster and selects an optimal node deployment MSU; and the scheduling weight and resource allocation are adaptively optimized through fast / slow feedback circulation of the monitoring module during operation. According to the method, the expandability of the system is ensured, meanwhile, the cross-node communication delay is remarkably reduced, and the task throughput is improved.

A system and method for adaptive ideation and innovation management using graph-based similarity computation and real-time social facilitation are disclosed, and the system may include a plurality of client devices operated by a human user and one or more AI Agents. The system may further include a computing platform running a Precompute Similarity Engine that may be configured to embed and index idea objects, precompute similarity clusters, and generate candidate merge or purge operations together with diversity and novelty signals; and running a Social Physics Engine that may be configured to monitor human user and machine participant (AI agent) interactions, compute group metrics, and issue digital facilitation interventions. The Precompute Similarity Engine and the Social Physics Engine may operate in an interdependent feedback loop such that diversity and novelty signals guide the generation of digital facilitation interventions that are communicated via the network to the plurality of client devices and the one or more AI Agents.

A method and system are disclosed for improving the accuracy, efficiency, and scalability of data interpretation and extraction of structured data from unstructured documents utilizing Large Language Models (LLMs). Applicable in finance, healthcare, legal, and government contexts, the disclosed invention addresses limitations of conventional Optical Character Recognition (OCR), machine learning, and LLM-based methods. In particular, the system and method incorporate feedback loops for continuous learning and leverage pre-processing, contextual tagging, customized prompt engineering, and post-processing to achieve robust data extraction. By integrating data enrichment techniques, the invention manages the inherent complexities of multilingual documents and evolving content standards.

The invention relates to a system cue word dynamic generation method and system based on multi-model feedback loop. The method comprises the following steps: (1) receiving an original cue word of a user; (2) calling an auxiliary large language model to perform semantic extension on the cue word of the user; (3) inputting each candidate system prompt word and a user prompt word combination into the main large language model; (4) performing multi-dimensional quantitative scoring on each preliminary answer through an evaluation model; (5) selecting an optimal system cue word according to the comprehensive score sequence, and triggering a main big model to generate a final answer; and (6) if all the preliminary answer scores are lower than a preset threshold value, starting a feedback loop to regenerate candidate system cue words. The method has the advantages that (1) the system cue word and the user cue word are strictly distinguished, and the intention purity of the user is ensured; and (2) the system prompt words are dynamically generated, and the adaptability is high.

The invention discloses an online low-voltage deduction method and system for a power distribution area based on collaborative driving modeling, and the method specifically comprises the steps: carrying out the deduction of a mechanism base layer through employing linear DistFlow, carrying out the embedded topology verification and power balance verification, and outputting basic voltage distribution; the data driving compensation layer learns spatial coupling and time dynamic characteristics of the voltage residual error by using GCN-LSTM, a PINN is adopted to embed power flow / physical constraint into network training, and a dynamic compensation amount is obtained through synthesis; the hybrid deduction engine calculates a self-adaptive weight and fuses the self-adaptive weight with the self-adaptive weight to obtain voltage prediction; the rolling deduction and incremental learning module continuously outputs a voltage situation map, and triggers online adjustment of parameters of a data driving compensation layer when an error exceeds a threshold; and the bidirectional feedback loop corrects the input deviation of the mechanism base layer and adjusts the data driving compensation layer. According to the invention, while the physical consistency is ensured, the precision and real-time response capability of voltage deduction are greatly improved.

The invention discloses a big data analysis and automatic evaluation system in project cost consultation, and relates to the technical field of big data analysis and evaluation. By constructing a unified data acquisition framework and a dynamic updating mechanism, efficient integration and real-time interaction of multi-party heterogeneous data are realized, data transparency and reliability are remarkably improved, a solid data basis is provided for project cost management, the system quickly and accurately optimizes various variables by utilizing comprehensive analysis and multivariable optimization technologies, and the system has a good application prospect. The system adapts to the dynamic environment, enhances the anti-risk capability of the project, ensures that the project cost management meets the real-time requirement and realizes the globally optimal solution, and in addition, through resource allocation strategy generation, view updating and decision recording and a feedback loop mechanism, the system realizes the generation and continuous improvement of a resource allocation strategy oriented to global benefit balance. The resource allocation efficiency and the long-term competitiveness of the project are improved, and powerful support is provided for sustainable development of urban construction.

The invention discloses an automatic equipment fault diagnosis and intelligent repair system oriented to complex working conditions, and belongs to the technical field of intelligent operation and maintenance of automatic equipment. The edge computing node is internally provided with a spatial-temporal feature separation convolutional network ST-FSCN and is used for processing multi-modal sensing data in real time; the cloud decision center runs a dynamic reinforcement repair strategy library and comprises a deep Q network based on a physical model, a meta-learning strategy optimizer MAML-RL and a security verification engine; the execution feedback loop is connected with the equipment PLC through an industrial bus, adjusts the operation parameters of the equipment in real time, and feeds back the execution effect to the self-evolution digital twin system; and the self-evolution digital twin system comprises a high-fidelity simulation model and a lightweight real-time model, and adopts a dual-channel updating mechanism. The method has the advantages that the composite fault recognition accuracy is improved, the average repair decision time is shortened, and accidental shutdown of equipment is reduced.

The method for generating mitigation policies may include receiving, by a large language model (LLM), a plurality of inputs including textual description of vulnerabilities, code with one or more vulnerability exploits, and meta-data; generating, by the LLM, API and application vulnerability mitigation policies based on the received inputs; generating, by a feedback loop automation engine (FLAE), feedback data based on the API and application vulnerability mitigation policies and the code with one or more vulnerability exploits; and updating, by the LLM, the API and application vulnerability mitigation policies based on the feedback data generated by the FLAE.

A log system prompts a foundation model to identify regular expressions that delineate single line logs in log files and incorporates user feedback for the regular expressions in a feedback loop as part of a hybrid parsing approach. The log system splits the log files into single line logs according to the regular expressions and augments the single line logs with classifications and values of metadata fields. The augmented single line logs are then clustered into hierarchical clusters with corresponding cluster patterns and the log system stores the augmented single line logs in a database indexed by the classifications and values of metadata fields and associated with corresponding clusters / cluster patterns. A presentation module accesses the database when responding to user queries for filtered single line logs and log correlations / root cause analysis.

The invention relates to the technical field of electric power management, and particularly discloses a multi-element load short-term prediction system for an electric power system, and the system obtains an actual operation data source and a meteorological data source of a power grid through a multi-source data collection platform, and carries out the preprocessing of the data sources. The method comprises the steps of extracting local fluctuation characteristics according to an actual operation data source and a meteorological data source of a power grid, fusing to generate a composite data sequence, designing a load trend sensing model based on the operation data sequence and the composite data sequence to analyze a power load recovery trend, responding to power grid dispatching according to a prediction result, and forming a feedback closed loop. The prediction and scheduling strategy is continuously corrected and optimized, internal and external multi-dimensional information is fused, and the system structure of short-term fluctuation and long-term trend is considered, so that the prediction precision is improved, and the robustness and the adaptive capacity of the system are remarkably enhanced.

Systems and methods for bias testing and remediation are disclosed. Decentralized Web Application Testing Systems use a Holochain framework to distribute testing workloads across Full Nodes and Lightning Nodes. A Holochain Node Management Application for configuring nodes, a UI Application creates test cases, and a Version Management System tracks changes. Test results are stored and analyzed in a Test Result Store, with a Bias Intelligence module detecting biases and generating additional test cases. A Consensus Algorithm validates test cases through decentralized nomination. An AI-Driven Defect Remediation System automates defect detection, root cause analysis, and remediation using AI modules. Machine learning algorithms identify patterns and anomalies, while NLP techniques generate code fixes. Predictive maintenance monitors application performance to preemptively address issues. A feedback loop mechanism continuously improves AI models through reinforcement learning. Together, these systems provide a holistic approach to web application testing and maintenance.

Systems and methods analyze data received from assets in a manufacturing environment, determine a first parameter is over or under a first predetermined threshold for an asset using a first machine learning model deployed on the manufacturing system, and generate an updated parameter value to adjust the asset. The systems and methods analyze a subset of sensor data in the manufacturing system using the first machine learning model, determine an actual accuracy rate of the first machine learning model based on an output by the first machine learning model from analysis of the subset of the sensor data, and based on determining that the actual accuracy rate is below a predefined threshold corresponding to a deployed accuracy rate, trigger a retraining of the first machine learning model using at least the subset of sensor data in the camera feed as training data.

The invention provides a code testing method and device, electronic equipment and a storage medium, and the method comprises the steps: responding to a to-be-tested code submission operation, and generating a target test case according to a to-be-tested abstract syntax tree and a to-be-tested demand document of a to-be-tested code; generating a target UI test script corresponding to the target test case according to interactive elements in the to-be-tested code user interface; and executing a test task constructed based on the target test case and the target UI test script to obtain a test result. The labor cost and the time overhead of a test link are remarkably reduced, a rapid feedback closed loop from code submission to test execution is constructed, and the software development efficiency and the quality guarantee capability are integrally improved.

An apparatus may include a set of modules and artificial intelligence models to detect a cyber incident, a simulator to simulate an actual cyber attack of the cyber incident on a network including physical devices being protected by the set of modules and artificial intelligence models; and a feedback loop between i) the set of modules and artificial intelligence models and ii) the simulator, during an ongoing detected cyber incident. An attack path modeling module is configured to feed details of the detected incident by a cyber threat module into an input module of the simulator, and to run one or more hypothetical simulations of that detected incident in order to predict and control an autonomous response to the detected incident. Any software instructions forming part of the set of modules, the artificial intelligence models, and the simulator are stored in an executable form in memories and executed by processors.

A method, a network device, and a non-transitory computer-readable storage medium are described in relation to a coordinated domain-based low latency, low loss, and scalable throughput (L4S) service. The service may enable end-to-end congestion control based on a predetermined order of domains, which are associated with L4S devices that support a packet flow, for which congestion control responses are to be provided. The service may provide independent feedback loops of state information from each domain and associated L4S device. The feedback loops may operate on different timescales. The service may coordinate congestion control among domains of an end-to-end path that support the packet flow and prevent fragmented or contentious congestion control responses between the domains.