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

2: Belief propagation is known to perform extremely well in many

3: practical statistical inference and learning problems using

4: graphical models, even in the presence of multiple loops. The use of

5: the belief propagation algorithm on graphical models with loops is

6: referred to as Loopy Belief Propagation (LBP). Various sufficient

7: conditions for convergence of LBP have been presented; however,

8: general necessary conditions for its convergence to a unique fixed

9: point remain unknown. Because the approximation of beliefs to true

10: marginal probabilities has been shown to relate to the convergence

11: of LBP, several methods have been explored whose aim is to obtain

12: distance bounds on beliefs when LBP fails to converge. In this

13: paper, we derive uniform and non-uniform error bounds on LBP, which

14: are tighter than existing ones in literature, and use these bounds

15: to study the dynamic behavior of the sum-product algorithm. We

16: subsequently use these bounds to derive sufficient conditions for

17: the convergence of the sum-product algorithm, and analyze the

18: relation between convergence of LBP and sparsity and

19: walk-summability of graphical models. We finally use the bounds

20: derived to investigate the accuracy of LBP, as well as the

21: scheduling priority in asynchronous LBP.

22: \end{abstract}

23: