All news
Digital Twin

Biases in end-to-end digital twins built from chained models

US & West ·

The work examines the statistical properties of a Monte Carlo scheme that forwards full predictive draws through a chain of separately fitted stage models. Two-stage analytical examples show persistent interval and coverage inflation and, with quadratic propagation, parameter-dependent mean bias; simulations of four- and eight-stage chains confirm pronounced distortions.

Why it matters

It warns that naively assembling a twin from stage models yields biased predictions and points to where correction is needed in end-to-end modelling.

Relevant to DT products

Original headline
Post-Inference Model Chaining: A Problem Statement and a Model-based Bootstrap for Training End-to-End Digital Twins
Read the original: arXiv API — полнотекстовый поиск "digital twin"

Translation and summary are machine-generated from the source. The full article is not republished; its rights belong to the source.

Related

Digital twinUS & West

Digital twin speeds up reinforcement learning for lab robots

TwinBridge builds digital twins from 3D Gaussian splatting and physics simulation to supply policy initialization and dense visual rewards for real-world online learning of a laboratory robot. Simulator task events and geometric progress yield rewards for pretraining, distilled into a visual reward model and combined with execution history during human-in-the-loop refinement.

arXiv API — полнотекстовый поиск "digital twin" ↗
Digital twinUS & West

Dataset for deciding between evidence gathering and action in control-system incidents

The authors present TE-GvA, a counterfactual decision dataset for fault-driven incident response on the revised Tennessee Eastman process: 9,908 decision points from 1,990 simulated incidents, each with 4–35 candidate actions labelled via branched rollouts. With labels computed under the true hidden cause, gathering evidence is the best choice in only 1.1–1.3% of decisions.

DTEDT.ForecastarXiv — Computational Engineering, Finance, and Science (cs.CE) ↗
Digital twinUS & West

Sparse stochastic twins can outperform exact digital twins

The study shows that keeping weakly dependent or auxiliary degrees of freedom raises the dimensionality of learning and control problems, so a sparse stochastic surrogate can beat an exact high-dimensional model. It introduces a method for dependence discovery at the level of the full conditional law, combining conditional generative modeling, Gaussian-process distillation and kernel mode decomposition.

DT.ForecastarXiv — Computational Engineering, Finance, and Science (cs.CE) ↗