1: \begin{abstract}

2: The purpose of this paper is to develop a 

3: practical strategy to accelerate Newton's method in 

4: the vicinity of singular points.

5: We present an adaptive safeguarding scheme 

6: with a tunable parameter, which we call 

7: adaptive $\gamma$-safeguarding, that 

8: one can use in tandem with Anderson acceleration

9: to improve the performance of Newton's 

10: method when solving 

11: problems at or near singular points. The key features of adaptive 

12: $\gamma$-safeguarding are that it converges locally 

13: for singular problems, and it can detect nonsingular problems 

14: automatically, 

15: in which case the Newton-Anderson 

16: iterates are scaled towards a standard Newton step. 

17: The result is a flexible algorithm that performs well for singular and nonsingular problems, and can recover convergence from both standard Newton and Newton-Anderson with the 

18: right parameter choice.

19: This leads to faster local convergence compared to both 

20: Newton's method, and Newton-Anderson without safeguarding, 

21: with effectively no additional computational cost.

22: We demonstrate three strategies 

23: one can use when implementing Newton-Anderson and 

24: $\gamma$-safeguarded Newton-Anderson to solve parameter-dependent

25: problems near singular points. For our benchmark problems, we  

26: take two parameter-dependent incompressible flow 

27: systems: flow 

28: in a channel and Rayleigh-B\'enard convection. 

29: \end{abstract}

30: