interpret

A natural experiment is an experiment (an AB-test if you will) that occurs by chance rather than conscious design. For example, two neighboring countries contemplate banning smoking in restaurants, but in one the bill fails with 49% of the vote, in the other the ban goes through with 51% of the vote. It's not perfect, but you could argue that these countries can now be used to estimate the effect of a smoking ban on mortality and health in a way that is almost as good (but not quite as good) as a randomized clinical trial, whereas you can't just compare two arbitrary countries with differential rates of smoking, because they might be different on so many other counts as well and there is no pre-intervention data to serve as a baseline. More broadly, ML does not really answer questions like "was this death caused by smoking, or rather by a

Uncertainty is absolutely key everywhere decision making has an impact on human lives. Detecting if we see an expresso or a cappucino, sure fine. But uncertainty on a CT scan would be extremely valuable just to avoid people (even trained professionals) falling into the common bias of trusting the output. Say the model outputs 90% chance of nothing, 10% chance of a deadly condition that needs to be investigated ASAP. Doctors are often overworked and tired and can miss it. Knowing where the certainties and uncertainties of the model lies would allow the doctor to compare that to its own training "Ah but it didn't at all judge based on X and Y" and limit decision machine based on "trusting the algorithms". This is a similar problem that rating system based on just a number of stars are facing, you need to read the comments and the why just t