Dies ist unsere neueste Auswahl an weltweiten Veröffentlichungen und Patenten in englischer Sprache zum Thema Neuronale Netze, aus vielen wissenschaftlichen Online-Zeitschriften, klassifiziert und fokussiert auf neuronale Netze, künstliche Neuronen, Epochen, neuronale Architektur, maschinelles Lernen, Deep Learning und Support Vector Machine.
Prevention of poor performance
Patent published on the 2026-05-21 in WO under Ref WO2026104713 by NOKIA TECHNOLOGIES OY [FI] (Fevold Jerediah [us], KovÁcs IstvÁn Zsolt [dk], Pantelidou Anna [fr], Tomala Malgorzata [pl], Hassan Sakira [fi], Wolfner GyÖrgy TamÁs [hu])
Abstract: Example embodiments of the present disclosure are directed to prevention of poor performance. A method comprises receiving, at the first apparatus from the second apparatus, a configuration for a counter associated with applicability determination information for a machine learning configuration; in accordance with receiving from the second apparatus, information indicating at least one cause for releasing a function configured by the machine learning configuration, incrementing the counter; and[...]
Our summary: This disclosure focuses on preventing poor performance in machine learning configurations. It involves receiving a configuration for a counter and information indicating causes for function release. The counter is incremented and reported back to indicate applicability based on its value.
performance prevention, machine learning, applicability determination, counter configuration
Patent
Physics-inspired machine learning for reliable production forecast in unconventional reservoirs
Patent published on the 2026-05-21 in US under Ref US20260141462 by CONOCOPHILLIPS CO [US] (Zhou Hui [us], Rincones Mazaruny [us], Montilla Lucybel [us], Orogbemi Olakunle A [us])
Abstract: Implementations described and claimed herein provide systems and methods for an innovative machine learning-driven approach, rooted in the fundamental physics of flow within fractured tight reservoirs for production forecasting of unconventional reservoirs. A first component of the method is to automatically analyze production data and generate characteristic attributes for linear flow and boundary-dominated flow. Following this, a Markov chain Monte Carlo process is utilized to integrate actual[...]
Our summary: The method utilizes physics-inspired machine learning for forecasting production in unconventional reservoirs. It analyzes production data to identify flow characteristics and employs a Markov chain Monte Carlo process for probabilistic modeling. A two-step machine learning model predicts future well performance based on flow regime characteristics.
machine learning, production forecasting, unconventional reservoirs, Markov chain Monte Carlo
Patent
Deep learning model for approximating inverse-matrix operations
Patent published on the 2026-05-21 in US under Ref US20260141285 by HEWLETT PACKARD ENTPR DEV LP [US] (Strenski David [us], Sukumar Sreenivas Rangan [us], Nanos Jordan [us])
Abstract: Systems and methods are provided for a model-centric approach that can be used to measure a computer s performance based on metrics obtained during and/or from training a machine learning (ML) model. Examples include building a training data set by generating first matrices and second matrices and deriving third matrices from the first and second matrices. Examples also include training, at a plurality of computer systems, a plurality of machine learning (ML) models by applying the first and thi[...]
Our summary: This content describes a deep learning model designed for inverse-matrix operations. It outlines a model-centric approach to assess computer performance using metrics from ML training. Additionally, it details the generation of training datasets and the application of ML algorithms to evaluate performance across multiple systems.
deep learning, inverse matrix, performance metrics, machine learning
Patent
Virtual reality pose estimation apparatus and method using haptic device
Patent published on the 2026-05-21 in WO under Ref WO2026106377 by UNIV NAT CHONNAM IND FOUND [KR] (Kim Myeong-jin [kr], Yong Han-bit [kr], Kim Hyeon-su [kr])
Abstract: The present invention relates to a virtual reality pose estimation apparatus and method using a haptic device. The apparatus comprises: an input unit for receiving, through a plurality of input devices, input point data for the estimation of a pose of a haptic device; a global feature extraction unit which inputs the input point data into a multilayer perceptron (MLP) model so as to extract point-wise features, and which extracts a global feature on the basis of the point-wise features; and a po[...]
Our summary: The invention details a virtual reality pose estimation system utilizing a haptic device. It includes an input unit for gathering data from multiple devices to estimate the haptic device s pose. A multilayer perceptron model extracts features from the input data, which are then processed by a hierarchical neural network for pose estimation.
virtual reality, pose estimation, haptic device, neural network
Patent
System and method for synchronous aggregation and amalgamation of occupational specifications by intelligent neural networking
Patent published on the 2026-05-21 in US under Ref US20260141350 by THOMPSON II DARRELL [US] (Thompson Ii Darrell [us])
Abstract: [0000] The invention discloses systems and methods for synchronous aggregation and amalgamation of occupational specifications by intelligent neural gathering of US Department of Labor Standard Occupational Classification, and Occupational Information Network standardized data descriptions. The system constructs standardized data configurations of job canon via program-global area data files and cumulative career historical data using system-global area data files, with the purpose of producing [...]
Our summary: The invention describes a system for synchronously aggregating occupational specifications using intelligent neural networking. It utilizes standardized data from the US Department of Labor to create job canon configurations. The system employs bi-directional recurrent neural networks for matching candidates to roles based on recruitment prerequisites.
neural networking, occupational specifications, data aggregation, machine learning
Patent
Multi-tenant system for well intervention candidate screening and ranking
Patent published on the 2026-05-21 in US under Ref US20260141291 by SCHLUMBERGER TECHNOLOGY CORP [US] (Paroha Abhay [us], Sinha Rajeev Ranjan [us], Vijayakumar Janaat [us], Yancey Casey [us], Von Niederhausern Eugene [us])
Abstract: [0000] The present disclosure relates to systems and methods for automatically identifying candidate wells for intervention opportunities in a field. The systems and methods use machine learning models to automate the data analysis to identify the candidate wells. The systems and methods provide insights for the candidate wells and recommendations for the intervention opportunities.[...]
Our summary: The system automates the identification of candidate wells for intervention using machine learning. It analyzes data to provide insights and recommendations for intervention opportunities. This multi-tenant approach enhances efficiency in well management.
multi-tenant, well intervention, machine learning, candidate screening
Patent
Federated spatiotemporal transformer for cross-facility carbon emission prediction in dynamic scheduling of low-carbon workshops
Published on 2026-05-18 by @OXFORD
Abstract: AbstractAmid the global shift toward low-carbon manufacturing, workshop-level dynamic scheduling has become a key approach for reducing carbon emissions due to its direct influence on production processes. However, most existing studies focus on single-workshop optimization and rarely explore emission reduction through cross-facility collaboration. Meanwhile, traditional carbon emission prediction methods struggle to build global carbon footprint models because of data silos, which limits the de[...]
Our summary: This study introduces a federated spatiotemporal transformer framework for predicting carbon emissions across multiple facilities. It addresses data silos and enhances prediction accuracy while maintaining data privacy. The framework integrates real-time forecasting with optimization algorithms to improve low-carbon scheduling strategies.
federated learning, spatiotemporal transformer, carbon emission prediction, low-carbon scheduling
Publication
Development strategy of rural low-carbon ecotourism under the background of big data and the Internet of Things
Published on 2026-05-12 by @OXFORD
Abstract: AbstractThrough the use of multimodal data, including behavioral, physiological, and textual sentiment inputs, this study presents a deep learning-based framework for sustainable rural ecotourism that monitors the mental health of employees working in rural ecotourism enterprises. The system exhibits excellent predictive performance (ROC-AUC = 0.945, F1-score = 0.913, and accuracy = 89.3%, equivalent to 0.893). It provides early warnings for stress, anxiety, and burnout up to 5.6 da[...]
Our summary: This study presents a deep learning framework for sustainable rural ecotourism that monitors employee mental health. It predicts stress and burnout earlier than existing methods, enhancing stakeholder collaboration. The framework aims to promote sustainable economic development in rural areas through efficient resource management.
rural ecotourism, big data, deep learning, mental health
Publication