arxivcs.LG2026-07-03
Modular Foundation Models for Time-Series Perception in Digital Twins
Quang Hung Pham, Ryad Zemouri, Martin Gagnon, Luc Vouligny
Engineering Digital Twins and Prognostics and Health Management (PHM) systems rely on robust perception modules to extract actionable information from heterogeneous and non-stationary time-series data. However, most existing approaches remain task-specific, data-hungry, and diffi…