Id: arxiv/2607.12613 Url: https://arxiv.org/abs/2607.12613 Tags: math.NA, math.OC Score: 0 Title: Tensor-Based Reduced-Order Modeling for Optimization-Based Inverse Problems Domain: math.NA Source: arxiv Authors: Sahidul Islam, Andreas Mang, Maxim Olshanskii Reviews: 0 Subject: math.NA Category: research Comments: 0 Disputed: false Division: math Mentions: 0 Discussed: 0 Published: 2026-07-14 Page: https://delphi.science/r/arxiv/2607.12613 We develop a tensor reduced-order modeling (TROM) framework for optimization-based inverse problems governed by parameter-dependent dynamical systems. The approach approximates the parameter-to-observation map directly in tensor-train (TT) format using TT-SVD or TT-Cross and integrates it into a regularized nonlinear least-squares formulation. Beyond accelerating forward evaluations, the low-rank tensor structure reformulates the inverse problem in reduced coordinates, assembles Gauss--Newton quantities without forming the full observation-space Jacobian, and minimizes the TROM objective…