1: \begin{abstract}

2: %There are increasing popularity and applications of Bayesian inference in the twenty-first century, but 

3: Although Bayesian inference is an immensely popular paradigm among a large segment of scientists including statisticians,

4: most of the applications consider the objective priors and need critical investigations \citep{Efron:2013}. 

5: And although it has several optimal properties, one major drawback of Bayesian inference is 

6: the lack of robustness against data contamination and model misspecification, 

7: which becomes pernicious in the use of objective priors. 

8: This paper presents the general formulation of a Bayes pseudo-posterior distribution yielding robust inference. 

9: Exponential convergence results related to the new pseudo-posterior and the corresponding Bayes estimators 

10: are established under the general parametric set-up and 

11: illustrations are provided for the independent stationary models and 

12: the independent non-homogenous models. 

13: For the first case, the discrete priors and the corresponding maximum posterior estimators are discussed 

14: with additional details. We further apply this new pseudo-posterior to propose robust versions of 

15: the Bayes predictive density estimators and the expected Bayes estimator for 

16: the fixed-design (normal) linear regression models; 

17: their properties are illustrated both theoretically as well as empirically.

18: 

19: %\keywords{Pseudo-Posterior \and Robustness \and Bayes Estimation \and Density Power Divergence}

20: \smallskip

21: \noindent

22: \textbf{Keywords:} Robust Bayes Estimator; Pseudo-Posterior; Density Power Divergence;

23: Exponential Convergence; Predictive Density Estimator; Bayesian Linear Regression.

24: \end{abstract}

25: