ForecastingControl Architecture for Fast Frequency Regulation with High Inverter-Based Resource PenetrationarXiv — Systems and Control (eess.SY) · 4hModellingNeural Network-Based Parameter Estimation for a Labour Market ABMarXiv — Multiagent Systems (cs.MA) · 4hForecastingMeteorology-Driven Causal Nowcasting of Landfill Fugitive Emissions for Proactive Public HealtharXiv — Computers and Society (cs.CY) · 4hDigital twinPalantir-Powered Digital Twin May Reshape Mercury Systems' Investment Casesimplywall.st · yesterdayDigital twinTwin: Playing an Unknown Game with a Test-Time Digital TwinarXiv API — полнотекстовый поиск "digital twin" · 2dModellingLamacon and Focus Group Combine Expertise in Optimization and Simulation ModelingCNews · 3dDigital twinEfficient Surrogate-Assisted Electromagnetic Parameter Calibration for 6G Digital Twin ChannelsarXiv API — полнотекстовый поиск "digital twin" · 4dDigital twinDigital Twin Satellite Networks: A Paradigm for Intelligent and Resilient OperationsarXiv API — полнотекстовый поиск "digital twin" · 4dModellingTopological Feasibility Guarantees for Differentiable Predictive ControlarXiv — Systems and Control (eess.SY) · 5dDigital twinOnline Data-Driven Reachability Analysis using Zonotopic Recursive Least SquaresarXiv — Systems and Control (eess.SY) · 5dDigital twinResults of Foresight Session on Cyber-Physical Systems Security PublishedspbIT · 4dDigital twinDigital Twin Networks for 6G: Architecture and ChallengesarXiv API — полнотекстовый поиск "digital twin" · 4d
Hierarchical Federated Transfer Learning in Digital Twin-Based Vehicular Networks
US & West ·
For digital twin-based vehicular networks (DT-VANET), a hierarchical federated transfer learning (HFTL) method is proposed to address data heterogeneity and sparsity among vehicles, improving global model accuracy. Algorithms for cloud server model update and intra-cluster federated transfer learning are developed.
Why it matters
The method improves prediction in distributed systems with heterogeneous data, which is significant for managing traffic flows and other large-scale systems.
Analysts discuss that building a digital twin of Mercury Systems' operations using Palantir technologies could improve transparency and management efficiency, potentially enhancing the company's investment appeal.
The Twin system builds an executable world model at test time for continual learning tasks like ARC-AGI-3 games. The model is constructed from simulation and interaction alone, without hand-engineering. Actions are validated in a twin world model, and mismatches are used to repair the model.
An SGWO-IM algorithm is proposed for calibrating material parameters in ray-tracing used for 6G digital twin channels. It combines Grey Wolf Optimizer with individual memory and an online surrogate model for candidate pre-screening, reducing computational cost and improving accuracy.
