PIS: A Generalized Physical Inversion Solver for Sparse Observations Using Diffusion Models

This research introduces a novel approach to solve physical inversion problems using set-conditioned diffusion models, potentially advancing the field of inverse problem solving. The paper's focus on sparse observations suggests an attempt to address real-world data limitations, which could be impactful.

• •Proposes a new solver, PIS, for physical inversion problems.
• •Utilizes set-conditioned diffusion models.
• •Addresses the challenge of sparse observations.