1: \begin{abstract}

2: 		Physics informed neural networks (PINNs) are deep learning based techniques for solving partial differential equations (PDEs)

3: 		encounted in computational science and engineering.

4: 		Guided by data and physical laws, PINNs find a neural network that approximates the solution to a system of PDEs.

5: 		Such a neural network is obtained by minimizing a loss function in which 

6: 		any prior knowledge of PDEs and data are encoded.

7: 		Despite its remarkable empirical success

8: 		in one, two or three dimensional problems, there is little theoretical justification for PINNs.

9: % 		In this paper, we aim to establish a mathematical foundation of the PINNs methodology.

10: 		

11: 		As the number of data grows, PINNs generate a sequence of minimizers which correspond to a sequence of neural networks.

12: 		We want to answer the question: Does the sequence of minimizers converge to the solution to the PDE? 

13: 		We consider two classes of PDEs: linear second-order elliptic and parabolic.

14: 		By adapting the Schauder approach and the maximum principle, 

15: 		we show that 

16: 		the sequence of minimizers strongly converges to the PDE solution in $C^0$.

17: 		Furthermore, we show that if each minimizer satisfies the initial/boundary conditions,

18: 		the convergence mode becomes $H^1$.

19: 		Computational examples are provided to illustrate our theoretical findings.

20: 		To the best of our knowledge, this is the first theoretical work that shows the consistency of PINNs.

21: 	\end{abstract}

22: