The Navier-Stokes equations describe how fluids move, but proving that their solutions remain smooth forever in three dimensions is one of the most famous unsolved problems in mathematics. One approach, developed by Constantin and Iyer, rewrites the fluid velocity not as a direct quantity but as an average over many randomly perturbed particle paths. Each particle is carried along by the fluid flow plus some random noise, and the fluid velocity at any point is recovered by averaging a certain quantity, called the Weber field, over all these random paths. This paper takes that probabilistic representation and carefully dissects its mathematical structure.
The central technical achievement is showing that, when you split the averaged Weber field into its mean part and its fluctuations around that mean, the fluctuations satisfy a clean, self-contained equation that does not require knowing anything about higher-order statistics. Normally in turbulence and stochastic systems, equations for averages involve unknowns at the next level of complexity, creating an infinite chain of unsolvable equations. Here that chain collapses because all the random particle paths are driven by the same underlying noise, which causes certain complicated terms to simplify dramatically into a standard diffusion operator. From this simplified equation the authors extract rigorous bounds on how spread out the fluctuations can be in space and time.
The paper also carefully maps out what this approach cannot do, at least not without new ideas. The authors construct explicit counterexamples showing that several natural strategies for controlling the solution fail in subtle ways. For instance, controlling the average of a deformation measure does not let you control the deformation of the average, and spatial smoothness of the mean field does not automatically imply the kind of decay in time that would be needed to close a regularity argument. The paper closes by stating precisely which mathematical steps remain open, making it a careful accounting of both progress and obstacles in this probabilistic approach to one of mathematics's hardest problems.