1: \begin{abstract}

2: We consider the constrained sampling problem where the goal is to sample from a target distribution on a constrained domain. 

3: We propose skew-reflected non-reversible Langevin dynamics (SRNLD), a continuous-time

4: stochastic differential equation with skew-reflected boundary. 

5: We obtain non-asymptotic convergence rate of SRNLD to the target distribution in both total variation and $1$-Wasserstein distances.

6: By breaking reversibility, we show that the convergence is faster than

7: the special case of the reversible dynamics.

8: Based on the discretization of SRNLD, we propose skew-reflected non-reversible Langevin Monte Carlo (SRNLMC), and obtain non-asymptotic 

9: discretization error from SRNLD, and convergence guarantees

10: to the target distribution in $1$-Wasserstein distance.

11: We show better performance guarantees than the projected Langevin Monte Carlo in the literature that is based on the reversible dynamics.

12: Numerical experiments are provided for both synthetic and real datasets to show efficiency of the proposed algorithms.

13: \end{abstract}

14: