1: \begin{abstract}

2: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%% 

3: We demonstrate that the use of asymptotic expansion as prior knowledge in the {\it ``deep BSDE solver"},

4: which is a deep learning method for high dimensional BSDEs proposed by Weinan E, Han \& Jentzen (2017),

5: drastically reduces the loss function and accelerates the speed of convergence.

6: We illustrate the technique and its implications by Bergman's model with different lending and borrowing rates, 

7: and a class of quadratic-growth BSDEs. We also present an extension of the deep BSDE solver for reflected BSDEs

8: using an American basket option as an example.

9: 

10: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

11: \end{abstract}