Abstract We introduce a novel algorithm for estimating optimal parameters of linearized assignment flows image labeling. An exact formula is derived the parameter gradient any loss function that constrained by linear system ODEs determining flow. show how to efficiently evaluate this using Krylov subspace and low-rank approximation. This enables us perform learning Riemannian descent in space, ...

Interactions growing slower than a certain exponential of the square scalar field are well behaved when evolved under functional renormalization group linearized around Gaussian fixed point. They satisfy properties usually taken for granted, and reproduce standard perturbative quantization. However, more challenging effects appear, interactions grow faster this. We show explicitly that firstly,...