1: \begin{abstract}

2:  In this work we present, analyze, and implement a class of Multi-Level Markov chain Monte Carlo (ML-MCMC) algorithms based on independent Metropolis-Hastings proposals for  Bayesian inverse problems. In this context, the likelihood function involves solving a complex differential model, which is then approximated on a sequence of increasingly accurate discretizations. The key point of this algorithm is to construct highly coupled Markov chains together with the standard Multi-level Monte Carlo argument to obtain a better cost-tolerance complexity than a single level MCMC algorithm. Our method extends the ideas of Dodwell, et al. ``A hierarchical multilevel Markov chain Monte Carlo algorithm with applications to uncertainty quantification in subsurface flow,'' \textit{SIAM/ASA Journal on Uncertainty Quantification 3.1 (2015): 1075-1108,} to a wider range of proposal distributions.  We present a thorough convergence analysis of the  ML-MCMC method proposed, and show, in particular, that (i) under some mild conditions on the (independent) proposals and the family of posteriors,  there exists a unique invariant probability measure for the coupled chains generated by our method, and (ii) that  such coupled chains are uniformly ergodic. We also generalize the cost-tolerance theorem of Dodwell et al., to our  wider class of ML-MCMC algorithms. Finally, we propose a self-tuning continuation-type ML-MCMC algorithm (C-ML-MCMC). The presented method is tested on an array of academic examples, where some of our theoretical results are numerically verified. These numerical experiments evidence how our extended ML-MCMC  method is  robust  when targeting some \emph{pathological} posteriors, for which some of the previously proposed ML-MCMC algorithms fail. 

3: \end{abstract}

4: