1: \begin{abstract}

2:   Fast Incremental Expectation Maximization (FIEM) is a version of the

3:   EM framework for large datasets.  In this paper, we first recast

4:   FIEM and other incremental EM type algorithms in the {\em Stochastic

5:     Approximation within EM} framework. Then, we provide nonasymptotic

6:   bounds for the convergence in expectation as a function of the

7:   number of examples $n$ and of the maximal number of iterations

8:   $\kmax$.  We propose two strategies for achieving an

9:   $\epsilon$-approximate stationary point, respectively with $\kmax =

10:   O(n^{2/3}/\epsilon)$ and $\kmax = O(\sqrt{n}/\epsilon^{3/2})$, both

11:   strategies relying on a random termination rule before $\kmax$ and

12:   on a constant step size in the Stochastic Approximation step. Our

13:   bounds provide some improvements on the literature. First, they

14:   allow $\kmax$ to scale as $\sqrt{n}$ which is better than $n^{2/3}$

15:   which was the best rate obtained so far; it is at the cost of a

16:   larger dependence upon the tolerance $\epsilon$, thus making this

17:   control relevant for small to medium accuracy with respect to the

18:   number of examples $n$. Second, for the $n^{2/3}$-rate, the

19:   numerical illustrations show that thanks to an optimized choice of

20:   the step size and of the bounds in terms of quantities

21:   characterizing the optimization problem at hand, our results design

22:   a less conservative choice of the step size and provide a better

23:   control of the convergence in expectation.

24: \end{abstract}

25: