9735 results about "Learning models" patented technology

The invention discloses a lightweight-based satellite landslide image intelligent detection method, device and equipment and a medium, and relates to the technical field of disaster detection, and the method comprises the steps: obtaining a whole-scene optical satellite image containing a landslide and a non-landslide region and landform auxiliary data; a dynamic segmentation strategy is adopted to carry out differential segmentation and standardized preprocessing on an image based on topographic data, and a standardized image is obtained. A target landslide image is screened through a double-layer machine learning model, the target image is input into an improved lightweight convolutional neural network for processing, an initial detection result is obtained, finally edge optimization and coordinate calibration are performed on the result, and an accurate landslide area detection result is output. The identification precision of the landslide image is effectively improved through dynamic segmentation and double-layer screening, the improved lightweight convolutional neural network realizes efficient detection in a low-resource environment, and the accuracy of a landslide detection result is further improved.

The invention discloses a bidding document intelligent analysis and structuring method and system based on a multi-modal knowledge graph. The method comprises the steps of S1, obtaining bidding document content and technical data related to constructional engineering; s2, performing knowledge extraction on the text in the bidding document content, and constructing a multi-modal knowledge graph; s3, performing semantic optimization on the multi-modal knowledge graph to obtain an updated multi-modal knowledge graph; s4, converting the multi-modal knowledge graph into corresponding feature vectors, and fusing the feature vectors; s5, inputting the multi-modal data into the constructed multi-modal deep learning model for processing, and outputting element information of bidding document structuring; s6, performing hierarchical division on bidding document contents according to the element information, and identifying contents of different hierarchies; and S7, constructing a rule base according to the obtained technical data, and performing matching verification on the standard knowledge in the knowledge graph through the rule base.

A machine learning model may be configured for training using an associated learning technique. A model configured for end-to-end backpropagation may adapted for associated learning by introducing functions for projecting hidden vectors and labels to a shared representation space and for reconstructing labels from representation vectors. An associated learning loss may be calculated at each layer, with the resulting gradients backpropagated locally through that layer rather than all layers. A reconstruction loss may be calculated using each layer's output including the predicted label. Training by associated learning may be parallelized (e.g., layer by layer) to yield efficiency gains. In addition, associated learning training may be more robust to training label errors. The resulting model may be used to, for example, predict data sequences in an autoregressive manner in which subsequent portions of the output data sequence are predicted in part based on previous predicted portions of the output data sequence.

The invention discloses an intelligent supply chain management system and method based on artificial intelligence and big data, and belongs to the technical field of supply chain management and artificial intelligence, and the method comprises the steps: obtaining a state data sequence of a supply chain object, extracting abnormal features, and forming an abnormal feature data sequence, obtaining a supply chain environment and operation parameter time sequence aligned in time and space; and jointly inputting the abnormal feature data sequence and the supply chain environment and operation parameter time sequence into a pre-trained multi-modal deep learning model for fusion analysis, and outputting one or more key supply chain parameters causing the abnormal state and quantized abnormal fluctuation information thereof, accurately associating the key parameters with the specific physical position or visual form of the abnormal state on the supply chain object, and finally generating an association map; according to the invention, full-link closed loop from data perception, intelligent analysis to root cause visualization is realized, and the intelligent level and fault processing efficiency of supply chain management are improved.

The disclosed concepts relate to training a machine learning model to provide help sessions during a video game. For instance, prior video game data from help sessions provided by human users can be filtered to obtain training data. Then, a machine learning model can be trained using approaches such as imitation learning, reinforcement learning, and / or tuning of a generative model to perform help sessions. Then, the trained machine learning model can be employed at inference time to provide help sessions to video game players.

The invention discloses a urinary calculus CT image automatic segmentation method based on deep learning, particularly relates to the technical field of medical image processing, and is used for solving the problem of low geometric fidelity of a segmentation result caused by hardening artifacts when an existing deep learning segmentation method is used for processing a high-density urinary calculus CT image. The method comprises the following steps: acquiring a urinary calculus CT image, performing initial segmentation by using a deep learning model to generate an initial calculus segmentation region, evaluating texture heterogeneity degree and identifying a hardening artifact risk region by analyzing feature value distribution of a structure tensor field, and positioning an artifact-causing source point based on a CT imaging projection geometric principle by reversely tracing a spatial position relation. According to the method, boundary distortion features are identified by analyzing CT value profile curve form distortion features and local boundary curvature singularity features, geometric correction is performed on corresponding boundaries in an initial stone segmentation region according to the boundary distortion features, a final stone segmentation region is obtained, and the geometric accuracy and reliability of a segmentation result are effectively improved.

The invention relates to the technical field of automation and intelligent operation and maintenance of a power system, in particular to a knowledge graph-fused reinforcement learning switching operation anti-error verification method, which systematically extracts a multi-dimensional anti-error rule covering an operation sequence, an equipment state and an electrical safety distance by constructing an operation ticket knowledge graph, and improves the accuracy of the operation ticket knowledge graph. The defect that a traditional single-station anti-error system is incomplete in rule coverage is overcome, meanwhile, in combination with deep mining of a reinforcement learning model on historical operation data, implicit anti-error rules can be automatically extracted, illegal scenes which are not covered by a traditional rule base are supplemented, overall-process and multi-level accurate verification of switching operation is achieved, and the verification efficiency is improved. And the risks of misoperation and missing detection are greatly reduced.

The invention discloses an intelligent control and optimization system and method for the whole process of a photovoltaic power station. The method comprises the following steps: constructing a component-level high-fidelity digital twin through multi-source heterogeneous data fusion and an attention enhancement deep learning model; injecting annotated historical fault data, carrying out joint training by combining a feature embedding layer and an adversarial generative network, and generating a digital mapping body with fault evolution deduction capability; a three-layer architecture is introduced, reinforcement learning and a multi-objective optimization algorithm are fused, and autonomous decisions of power station level power output maximization, equipment loss minimization and maintenance cost optimization are realized in a virtual simulation environment constructed by a digital mapping body. According to the method, accurate dynamic mapping of the physical entity and the virtual space, fault pre-judgment and deduction and closed-loop control of global autonomous optimization are realized, the intelligent level and the full-life-cycle economy of the photovoltaic power station are improved, and the problems of insufficient fidelity, lack of fault deduction capability and low global optimization efficiency in the prior art are solved.

The invention relates to the technical field of block chains, in particular to a smart contract vulnerability detection method and device based on multi-modal features, and the method mainly comprises the steps: training a meta-learning model in a dynamic adaptation module, and adjusting the global parameters of a modal feature extraction module, a dynamic gating fusion module and a classifier through the meta-learning model, the dynamic adaptation module comprises a meta-learning model constructed based on an MAML framework, and is used for optimizing global parameters of each module according to vulnerability features learned in pre-training; and inputting the multi-modal fusion feature vector into a classifier, and generating and outputting a vulnerability detection result of the smart contract. According to the method, known vulnerabilities can be accurately detected by fusing multi-modal features, and novel vulnerabilities can be rapidly adapted and detected.

A machine learning method where, in a first step, a first (general) machine learning model is trained using a first training dataset including unlabelled optical fibre sensing data. Then, in a second step, a transfer learning process is applied to adapt or fine-tune the first machine learning model to a more specific application (e.g. to perform a specific type of detection or classification). Due to the large volumes of optical fibre sensing data available, the first machine learning model may provide a general machine learning model which has a high level of generality and is highly adaptable.

The invention relates to the technical field of transformer substation intelligent inspection, in particular to a meta-learning-based small-sample transformer substation equipment state adaptive inspection system, which comprises a state acquisition module for acquiring the current feature vector and environmental parameter data of a target node; the drift detection module is used for comparing environment parameters to judge data drift and dynamically adjusting a confidence coefficient threshold value; the risk assessment module inputs the feature data into a meta-learning model to output an initial risk probability, and generates an effective risk probability based on threshold filtering; the blind area measurement module is used for acquiring unobserved nodes and calculating system state blind area entropy; the scheduling decision-making module is used for comparing the blind area entropy with a threshold value and generating an entropy reduction bottom instruction or a self-adaptive routing inspection distribution instruction; the strategy updating module is used for extracting an actual inspection result and feeding back to the model for parameter updating; according to the invention, the scheduling difficulty when the resources are limited is solved, and the self-adaptive capability of the system under different environment interferences is improved.

In various examples, systems and methods are disclosed relating to displaying images in chatbot / NPC / virtual agent / digital avatar / etc. responses. A system can identify text corresponding to an image in an electronic document and can store a representation of the text in association with an identifier of the image. The system can receive an input prompt for a machine-learning model. The system can generate a response to the input prompt using the machine-learning model. The response can include the image responsive to identifying the representation of the text using a searching function and an output of the machine-learning model.

Systems and methods for searching using machine-learned model-generated outputs can provide a user with a medium for generating synthetic images depicting synthetic environments that can then be matched to a real world example. The systems and methods can include obtaining a search query, which can be utilized to generate a prompt input that can be processed by an image generation model to generate a plurality of model-generated images. A selection can then be received that selects a particular model-generated image to utilize to query a database.

The invention relates to the technical field of injection molding pressure control, in particular to a multi-stage pressure closed-loop compensation system and method for high-precision double-color injection molding, and the system comprises an interface gradient positioning module, a flow resistance state recognition module, a fitting function reconstruction module, a time sequence deviation triggering module and a gain reverse correction module. According to the method, the high-density measuring point data is collected, the interface radial pressure gradient change is extracted through the finite difference method, the stably-arranged nodes at the material intersection can be accurately locked, and the flow resistance state of the flow channel is judged according to the slope difference value of the screw speed and the pressure intensity change. A linear boosting or index slow increasing strategy is automatically switched to reconstruct a pressure function form, local high-pressure accumulation characteristics of a product sensitive area are accurately matched, the coincidence degree of a time window drift trend and a slope sudden change window is monitored, historical displacement and pressure errors are converted into backstepping factors through an iterative learning model, and the backstepping factors are calculated. And a driving gain curve is initialized and revised, so that forming deviation caused by thermal drift of equipment is eliminated.

The invention discloses an aero-engine residual life prediction method based on multi-modal deep learning, and relates to the field of aero-engine prediction and health management. The method comprises the following steps: acquiring and preprocessing multi-sensor time sequence data; constructing a degradation sensitive feature set through multi-scale analysis of a time domain, a frequency domain and a time-frequency domain; constructing a multi-modal deep learning model comprising an original data processing module and a multi-scale feature processing module, and introducing an attention mechanism and an uncertainty quantization module into the model; the model is subjected to lightweight processing to support embedded deployment. According to the method, multi-scale features and multi-modal deep learning are fused, the uncertainty quantification capability is achieved, high-precision and interpretable residual life prediction with uncertainty quantification is achieved, and reliable support is provided for engine maintenance decision making.

Certain aspects provide techniques and apparatus for executing queries in a computing system using machine learning models. An example method generally includes receiving a plan to satisfy a request in the computing system and event log data associated with execution of the plan. The plan generally specifies a first plurality of function calls at a first level of granularity. Using a plan refinement machine learning model, a refined plan is generated when the event log data indicates that execution of the generated plan results in one or more execution errors and the one or more execution errors are solvable. Generally, the refined plan specifies a second plurality of function calls at a second level of granularity, the second level of granularity being finer than the first level of granularity.

Predicting a future state, such as a future position and / or orientation (i.e., pose), of an object may comprise classifying, by a first machine-learned model, a lane the object may occupy and classifying, by a second machine-learned model, a target pose the object may occupy. A third machine-learned model may determine an offset from the target pose that may be used to determine a predicted (future) pose of the object by applying the offset to the target pose.

Embodiments provide a user device, UE, of a wireless communication network, the wireless communication network using one or more Artificial Intelligence / Machine Learning, AI / ML, models for one or more use cases, wherein the UE is configured or preconfigured with a plurality of AI / ML models for performing one or more certain operations, and wherein, dependent on one or more criteria, for performing the one or more certain operations, the UE is toswitch from a first AI / ML model to a second AI / ML model, ordeactivate one or more of the plurality of AI / ML models, orswitch from a current operation mode to a new operation mode.

Techniques for complex code modification via chain of thought prompting are described. A modification service obtains code to be modified according to a modification goal, and can decompose the modification goal, via use of a database or a machine learning model, to identify a set of modification steps. The set of modification steps may be organized according to a chain of thoughts, tree of thoughts, or graph of thoughts. The modification service can execute the set of modification steps via use of a machine learning model to yield modified code, which can be returned to a user.

The invention discloses a network space security intelligent monitoring and analysis system, and relates to the technical field of network security monitoring and analysis, and the system comprises a multi-source data collection module which collects multi-dimensional data in a full-link manner, and the collection frequency is dynamically adjusted along with a network load; the data preprocessing module cleans the fused data and generates a standardized analysis data set; the AI intelligent risk identification module identifies various safety risks in real time through a mixed deep learning model; the real-time response processing module starts differential processing according to a three-level mechanism; the threat traceability analysis module traces an attack link and generates a report; the security situation visualization module displays the security state in multiple dimensions; the data encryption storage module encrypts and protects data and performs double backup; and the system self-optimization module dynamically optimizes the strategy through incremental learning. The method is accurate in risk identification, timely in response processing, reliable in traceability and evidence storage, and efficient in cross-domain cooperation; terminal protection and third-party access control are enhanced, and network space security and stable service operation are comprehensively guaranteed.

Methods and systems are described for analyzing images. One or more machine learning models may be trained based on a plurality of images. The one or more machine learning models may comprise a model representing a feature in a scene. The one or more machine learning models may be trained to map input image coordinates to vectors of spline control points. Images may be reconstructed removing the feature from the scene.

The invention discloses a dynamic airspace gridding management method and system for low-altitude economy, and belongs to the technical field of unmanned aerial vehicle traffic management, and the method comprises the steps: collecting airspace state data in real time through a multi-source sensing device, and constructing a four-dimensional space-time grid model; generating a four-dimensional space-time grid with a block chain hash code by fusing meteorological data, an airspace control rule and a real-time flight demand; receiving a space-time grid use request submitted by the aircraft through the smart contract, and calculating an optimal grid allocation scheme based on a deep reinforcement learning model; and the edge computing node executes local track prediction, issues a navigation instruction to the aircraft through the distributed account book, monitors a grid occupation state in real time, and triggers a dynamic grid recombination mechanism when sudden conflicts are detected. According to the method, the rigid constraint of static airspace division can be broken through, the cooperative conflict of multiple aircrafts is eliminated, and the marketization configuration of airspace resources is realized.

This disclosure describes a framework for performing user-requested tasks automatically across an interactive interface using various types of machine learning models. Specifically, this disclosure outlines and describes a task execution system that utilizes a generative artificial intelligence (AI) action model and retrieval-augmented generation (RAG) to complete user-requested actions across an interactive interface. The task execution system solves many of the current limitations of LAMs by using a generative AI action model to determine a session plan, which includes a set of actions for accomplishing stages of the actionable task across the interactive interface, obtaining visual context information of each interactive interface segment, integrates RAG results to improve the accuracy of both the session plan and individual actions, and self-corrects when faced with unexpected obstacles.

A system and method are provided for predicting battery-performance information (e.g., remaining useful life (RUL), state of health (SOH), and / or state of charge (SOC)) for battery based on cycling data. For example, the battery-performance information can be predicted using machine learning (ML) models that predict battery-performance information for the battery based on cycling data and electrodynamic parameters (EDPs) that are generated either by calculating the EDPs using probing waveform data or predicting the EDPs from cycling data. The ML model can have been trained using results from a physics-based model when calculating the loss function used for training the ML model.

Systems and methods are disclosed for real-time anomaly prediction using near real-time data. The invention addresses delays in telemetry data collection from infrastructure components, by collecting metrics and logging this data in real-time. Extracted logged data undergoes initial analysis to identify patterns and anomalies, followed by cleaning to remove noise and errors. Feature engineering enhances the data, creating or modifying features to improve machine learning model performance. The system calculates weighted means of previous data values and computes first and second-order differences to capture immediate changes and trends. These calculations adjust the extrapolated value to accurately reflect current conditions. The adjusted data is integrated into the dataset and validated. The validated data trains and tests a machine learning model, which is then finalized and deployed for real-time anomaly detection. This system ensures accurate and timely anomaly prediction, enabling automated incident response to maintain the reliability and performance of infrastructure components.

The invention provides a system and method for managing the lifecycle of machine learning models, from development to deployment and ongoing operation, across various environments including on-premises, cloud, and hybrid infrastructures. The system features a model build platform for data processing, feature generation, model development, training, and hyperparameter tuning. A model analytics engine extracts metadata, performs complexity analysis, and generates configuration files specifying environment settings and resource needs. A secure model repository enables version-controlled storage, while a deployment platform retrieves, validates, and deploys models in containerized environments like OpenShift or Kubernetes. The platform dynamically allocates resources, supports real-time and batch scoring, and monitors model performance with guardrails. Customizable agents provide real-time feedback and automated optimization, and the system can securely decommission models while maintaining detailed lifecycle records. The invention enhances the efficiency, security, and scalability of machine learning operations with continuous performance improvement and compliance automation.

The present invention discloses an intelligent control method for sewage treatment equipment at dry bulk cargo terminals, relating to the technical field of sewage treatment equipment control, and is used to address the problem of poor control of sewage treatment equipment at terminals. The method comprises following steps: installing multiple types of sensors at key locations on dry bulk cargo terminals, utilizing edge computing nodes for real-time data acquisition and preprocessing, combining historical and temporal features with machine learning models to classify sewage types, achieving efficient dynamic adjustment of sewage treatment equipment operation parameters, then using weighted voting and confidence assessment to integrate multiple classification results to ensure optimal treatment results, and analyzing the actual sewage treatment situation to optimize equipment control continuously, thereby preventing failures, extending equipment life, and improving sewage treatment effectiveness.

A software package is received and unpacked into multiple components comprising plural functions. Each function is lifted from machine code into static single-assignment intermediate representation and tokenized to produce semantics-preserving embeddings. Intermediate-representation data-flow features are extracted, including detection of constant static variables on a stack, stack reaching definitions, uninitialized variables, and intra-procedural aliases. For each component, the embeddings and features are input to a machine-learning model trained on semantic properties derived from a corpus of software packages to generate a software supply chain risk level. Data characterizing the risk level is provided to a consuming application. When the risk level satisfies a remediation criterion, a remediation action is initiated, including generation of a source-code patch recommendation for an identified root-cause function, insertion of a runtime guard into the component, or issuance of a security advisory for distribution to a security operations dashboard.

The invention discloses a power distribution network transient characteristic prediction method based on supervised learning, and relates to the technical field of power distribution network state prediction, and the method comprises the steps: collecting historical operation data through a power distribution network monitoring system, carrying out the data preprocessing, and obtaining standardized multi-dimensional time series data; carrying out transient feature extraction, constructing a high-dimensional feature set, and carrying out feature dimension reduction according to a transient event tag to generate a feature subset; inputting the feature subset into a mixed supervised learning model of a gradient boosting decision tree GBDT and a long short-term memory network LSTM for joint training to obtain a transient feature prediction result; and calculating a root-mean-square error according to the transient characteristic prediction result and the real-time monitoring observation value of the power distribution network, and dynamically adjusting hyper-parameters of the supervised learning model based on a Bayesian optimization algorithm. According to the method, the detection accuracy can be improved, the calculation complexity can be reduced, and the discrimination capability and the time sequence prediction capability of the model are considered.

The invention relates to the technical field of medical simulation training, in particular to an ultrasonic operation intelligent training method and device based on multiple modalities, and the method comprises the steps: obtaining real-time six-degree-of-freedom pose data of an ultrasonic probe held by an operator, and generating an ultrasonic image in real time through a first deep learning model in cooperation with scene parameters; wherein the first deep learning model is trained to learn and establish a continuous mapping relation from an ultrasonic probe pose space to an ultrasonic image space; performing section classification on the ultrasound image by using a second deep learning model, and determining deviation information for a non-standard section; and based on the real-time six-degree-of-freedom pose data and a division result of a preset standard section and an ultrasonic probe pose, generating visual guide information for guiding an operator to adjust the probe pose, and displaying the visual guide information. In this way, the problems that an existing virtual training image is discontinuous and guiding is inaccurate are solved, and meanwhile the training efficiency and the reality sense of ultrasonic operation are remarkably improved.