1: \begin{abstract}%   <- trailing '%' for backward compatibility of .sty file

2: We study robust PCA for the fully observed setting, which is about separating a low rank matrix $\BL$ and a sparse matrix $\BS$ from their sum $\BD=\BL+\BS$. In this paper, a new algorithm, dubbed accelerated alternating projections, is introduced  for robust PCA % which accelerates existing alternating projections proposed in \citep{netrapalli2014non}.  %Let $\bm{L}_k$ and $\bm{S}_k$ be the current estimates of the low rank matrix and the sparse matrix, respectively. The algorithm achieves significant acceleration by first projecting $\BD-\BS_k$ onto a low dimensional subspace before obtaining the new estimate of $\BL$ via truncated SVD. 

3: which significantly improves the computational efficiency of the existing alternating projections proposed in \citep{netrapalli2014non} when updating the low rank factor. The acceleration is achieved by first projecting a matrix onto some low dimensional subspace before obtaining a new estimate of the low rank matrix via truncated SVD.

4: Exact recovery guarantee has been established which shows linear convergence of the proposed algorithm. Empirical performance evaluations establish the advantage of our algorithm over other state-of-the-art algorithms for robust PCA.\end{abstract}

5: