1.In 2001 Leo Breiman published his article “Statistical Modeling: The Two Cultures” in which he opines the idea that there are two core approaches (or “cultures”) in statistical modeling.  In more modern language this refers to approaches which assume that data come from a known distribution and others that believe the data can be simulated but which lack a known likelihood function, now referred to as generative models.  The application of machine learning to develop the generative model is often held up as the most appropriate option.  However, likelihood-free methods, specifically Approximate Bayesian Computation (ABC) is a lesser known alternative approach that has been actively developed over the last 10 years. 

Explain what Approximate Bayesian Computation is and what advantages it offers when a likelihood is unknown or intractable. 

Review the  benefits and its limitations of ABC, especially in the context of current data science.  

Finally, describe the link between machine learning and ABC and the potential points on conflict or consternation.