1: \begin{abstract}

2: 	

3: In this work, we consider solving optimization problems with a stochastic objective and deterministic equality constraints. We propose a Trust-Region Sequential Quadratic Programming method to find both \textit{first- and second-order} stationary points. Our method utilizes a random model to represent the objective function, which is constructed from stochastic observations of the objective and is designed to satisfy proper adaptive accuracy conditions with a high but fixed probability. To converge to first-order stationary points, our method computes a gradient step in each iteration defined by minimizing a quadratic approximation of the objective subject to a (relaxed) linear approximation of the problem constraints and a trust-region constraint.

4: To converge to second-order stationary points, our method additionally computes an eigen step to explore the negative curvature of the reduced Hessian matrix, as well as a second-order correction step to address the potential Maratos effect, which arises due to the~\mbox{nonlinearity}~of~the~problem constraints.~Such an effect may impede the method from moving away from saddle points.~Both~\mbox{gradient}~and~eigen~step~\mbox{computations}~\mbox{leverage}~a~novel~\textit{parameter-free}~decomposition of the step and the trust-region radius, accounting for the proportions among~the~feasibility residual, optimality residual, and negative curvature. We establish \textit{global almost sure} first- and second-order convergence guarantees for our method, and present computational \mbox{results}~on~CUTEst 

5: problems, regression problems, and saddle-point problems to demonstrate its superiority over~existing line-search-based stochastic methods.

6: 

7: \end{abstract}

8: