12 results about "Automated learning" patented technology

The invention provides a logging analysis model construction method and system based on multilevel regional feature association, and belongs to the technical field of oil-gas exploration artificial intelligence. In order to solve the problems of high cost, poor generalization and the like of a conventional method for performing data correction by using a standard well, the method comprises the following steps of: firstly, constructing multi-level geographic region embedding according to well location hierarchical codes, and capturing regional geologic features; then, constructing a regional relation graph based on well location space coordinates; and finally, aggregating features of adjacent regions through a graph neural network, and fusing multi-level region features and logging data through gating weight. According to the method, automatic learning of regional features is realized through a technology of performing gating weight fusion on the logging data by using the multistage regional nesting embedding module, the graph neural network spatial modeling module and the logging data, dependence and cost on standard well correction are removed, and cross-regional prediction precision is remarkably improved by establishing the regional features.

We provide the system. [Solution] A means of collecting surrounding information using various detection devices, An automated learning model means for analyzing the aforementioned information and detecting anomalies, A means for performing an evaluation based on anomaly detection and generating an alarm, A means of controlling mobile devices to patrol the site, A means of evaluating and optimizing the operation record of the entire system, A means to enable real-time confirmation of the aforementioned anomaly from a mobile device, A system that includes this.

The present disclosure provides a system for automated learning for a digital worker comprising one or more processors and a non-transitory machine-readable medium storing instructions. The instructions cause the processors to receive a video recording captured by a user demonstrating steps for executing a specific task, generate a raw instructions file comprising recorded steps and user input events, and process the video recording and raw instructions file using a machine learning module to generate a processed instructions file adapted to the task context. Processing comprises analyzing video frames to detect user interface elements and correlating detected elements with user input events. The operations further comprise validating the processed instructions file, assigning the specific task to a digital worker configured to execute automated tasks according to the processed instructions file, and executing the specific task by the digital worker.

This application provides an enterprise-level financial big data model platform, relating to the fields of artificial intelligence and fintech. The platform includes: a model layer for collaboratively integrating the capabilities of models with different modalities and parameter quantities, providing model capabilities to the component layer and application scenario layer; a component layer for encapsulating model capabilities into standardized components; and an application scenario layer for calling standardized components corresponding to task requirements based on financial business scenarios, combined with multimodal data, to meet the task requirements of the financial business scenarios. The method of this application improves the development efficiency of financial applications, reduces development costs, and shortens the development cycle by integrating the interfaces of different models through the componentization of big and small model capabilities. Through an automated learning mechanism, it continuously optimizes model capabilities based on actual business feedback, achieving continuous model optimization, enhancing the intelligence of financial applications, and reducing the labor and maintenance costs of commercial banks.

A computer implemented system and process to determine a model for a domain includes identifying a schema that defines a possible causal element of a particular type of behavior. One or more concepts are determined, as well as one or more sub-concepts for each concept, where each concept and sub-concept are associated with a logical relationship. Multiple models are determined from the one or more concepts and the one or more sub-concepts. The multiple models may be calibrated using representative data collected from a real-world source. An optimal model is determined amongst a plurality of calibrated models.

A method for determining by prediction, using automated learning of the machine learning type, time ranges during which a radio interface of a wireless access point of a communication network can be switched off or put on standby, for energy-saving purposes.

The invention relates to the field of machine vision, in particular to an on-line cosmetic contact lens printing ink dot detection method and system based on meta learning, and aims to realize rapid detection and quality judgment of ink dots of contact lenses with different patterns by adopting a Few-ShotCounting model in combination with a meta learning mechanism and through on-line training of a small number of template images. According to the system, through area division, ink dot counting and rule judgment, ink dots in an optical area and standard-exceeding ink dots on the edges of a printing area and the optical area are accurately recognized. The invention relates to a machine vision detection technology and a signal processing technology, and aims to further utilize a contact lens image collected by a high-resolution camera, automatically learn and identify the image, and judge whether a corresponding product is normal or not, so as to achieve the purposes of reducing labor cost, saving time and improving the production yield.

The invention relates to a method for sorting articles in a plurality of sorting devices, and a computer readable storage medium storing instructions for implementing the method, in which the method comprises the following steps: sorting (S1) a stream of articles (2) at a first sorting device (1) into at least one sorted portion (3) and optionally at least one unsorted portion (4); capturing (S2) a set of partial images; marking (S3) the partial image with a partial label; at least one classifier configured to correlate the item with the portion is constructed (S4), the at least one classifier being a machine learning model. Further, at the second sorting device (10), capturing (S5) an image of the items in the item stream (12); sorting of the articles in the stream of articles (12) into at least one sorted portion (13) and optionally at least one unsorted portion (14) is controlled (S6) using the images captured at the second sorting device (10) and at least one classifier.

The invention relates to an artificial intelligence-based data analysis system for translating and comparing file content. The system comprises: - a reading module configured to receive and process files in various formats; - a language conversion module configured to analyze and translate file content into simple and understandable language; - a module for comparing the converted data with that of other business entities; - a decision engine configured to receive information from the language conversion and comparison modules and to provide suggestions and strategies based on the results of the analyses performed; - a learning unit configured to undertake automated learning based on new data and new file formats, while continuously improving its performance and its ability to adapt to different business needs.Figure for the abridged version: 1.

The invention provides a research and judgment strategy optimization method and device based on a large model, equipment and a storage medium, and relates to the technical field of computers, in particular to the technical field of large language models, intelligent agent technologies and network security. According to the specific implementation scheme, unsupervised clustering is carried out on difference samples of safety alarm research and judgment to form at least one sample cluster containing similar features; when any sample cluster meets a preset triggering condition, calling a large language model to carry out attribution analysis on the sample clusters so as to infer potential reasons causing research and judgment differences; based on the potential reasons, generating a structured new research and judgment strategy; and optimizing the strategy library based on the new research and judgment strategy. According to the scheme disclosed by the invention, self-reflection is carried out on the misinformation of clustering through the large model, the optimization strategy is automatically generated, and an intelligent and automatic learning closed loop of the research and judgment process is constructed, so that the research and judgment accuracy is continuously improved, and the operation cost of manual analysis and strategy iteration is remarkably reduced.

A subject knowledge multi-hop question and answer method based on structured planning and reflection reasoning comprises the steps that firstly, for input subject questions, a structured planning module is adopted to conduct decomposition and subtask modeling on complex questions, and a multi-stage reasoning path conforming to a subject knowledge system is generated; secondly, dynamically retrieving related knowledge contents from teaching materials and subject knowledge bases, and constructing a structural chemistry knowledge context expressed in a triple form in combination with an information extraction technology driven by a thinking chain; and finally, introducing an inverse reasoning module, and performing consistency analysis and self-adaptive correction on an intermediate reasoning result, thereby realizing accurate subject knowledge questions and answers. According to the method, controllability of the reasoning process in subject knowledge question answering is achieved through structured planning, correctness and robustness of a reasoning chain are improved through reflection reasoning, the understanding depth of textbook knowledge and question answering performance are improved in combination with structured knowledge contexts, and the method is suitable for scenes such as intelligent teaching tutoring and automatic learning evaluation.

The present disclosure relates to the technical field of intelligent model, and particularly relates to a model construction method and device, equipment and a storage medium. The method comprises: acquiring spatio-temporal data, and acquiring task parameters required for model construction; based on the task parameters, performing standardization processing on the spatio-temporal data to obtain spatio-temporal graph data expressed in a standardized form; based on the spatio-temporal graph data and the task parameters, constructing a search space corresponding to a target spatio-temporal prediction model; based on the search space and the spatio-temporal graph data, performing network architecture search and network parameter optimization corresponding to the target spatio-temporal prediction model to obtain the target spatio-temporal prediction model. The present disclosure is used to solve the defect that the spatio-temporal data cannot be directly applied to the automatic learning in the spatio-temporal graph modeling task in the prior art, and realizes the process of automatically constructing the target spatio-temporal prediction model based on the spatio-temporal data.