Tensor Field Models

arXiv:2608.18808v1 Announce Type: new Abstract: This paper introduces Tensor Field Models (TFMs), realization-level Mathematical Structures in which a learned Operator maps a product of admissible component-section families to a prescribed family of time-dependent tangent sections on a Generative State Manifold. Analytic and dynamical restrictions are encoded through the choice of admissible families rather than imposed by the root definition. Constructed, component-separable, and Tensor Bundle TFMs provide structured refinements of this common object. In the conditional realizations considered here, a structured condition c=(c_1,ldots,c_n) is mapped componentwise to a reusable collection mathbf H_c=(H_{c_1}^{(1)},ldots,H_{c_n}^{(n)}). In the architectures evaluated here, the component representations remain distinct and are combined only by the Field Operator to produce the generated Vector Field. All learned models are trained using Flow Matching. Experiments show that TFMs can improve performance and that amortized sampling enabled by reusable condition representations can accelerate generation.

Related

• On the Approximation Complexity of Matrix Product Operator Born Machines
• Tensor Train Diffusion: Leveraging Low-Rank Structures for High-Dimensional Score-Based Sampling
• DeepRitzSplit Neural Operator for Phase-Field Models via Energy Splitting
• Kernel Neural Operators (KNOs) for Scalable, Memory-efficient, Geometrically-flexible Operator Learning
• Multi-stage neural operator learning with application for convolutions

Source: arXiv cs.LG | 2026-08-20