Sobolev improving for averages over curves in $\mathbf{R^4}$

Abstract: We study $L^p$-Sobolev improving for averaging operators $A_{\gamma}$ given by convolution with a compactly supported smooth density $\mu_{\gamma}$ on a non-degenerate curve. In particular, in 4 dimensions we show that $A_{\gamma}$ maps $L^p(\mathbb{R}^4)$ the Sobolev space $L^p_{1/p}(\mathbb{R}^4)$ for all $6 < p < \infty$. This implies the complete optimal range of $L^p$-Sobolev estimates, except possibly for certain endpoint cases. The proof relies on decoupling inequalities for a family of cones which decompose the wave front set of $\mu_{\gamma}$. In higher dimensions, a new non-trivial necessary condition for $L^p(\mathbb{R}^n) \to L^p_{1/p}(\mathbb{R}^n)$ boundedness is obtained, which motivates a conjectural range of estimates.