The most surprising bit of information I learned during the Prediction X interview with Professor Ned Hall are some of the negative implications of machine learning. Although I’ve often heard about the benefits of machine learning in media articles, it was insightful to learn that one of the negative elements of machine learning is that it can de-emphasize explanations. It was interesting to learn from Hall’s view that the final prediction is not the only benefit of the traditional framework because the understanding we receive from explanations is just as valuable. I agreed with Professor Goodman’s statement that we should avoid using machine learning blindly. In addition to separating what is important from what is convenient, I think scientists could also distinguish between the predictions that actually further a field of study versus simply add extraneous information.
I was also intrigued by Hall’s comment that historical and religious influences can impact the type of explanations that are offered in the prediction framework. It seems to me that in the prediction framework, these historical and religious factors could play a role in both the observation, data, and explanation categories. I would be interested in learning more about the concept of the “real” scientific method and how predictions elad to experiments which lead to an eventual stopping point.
Here is the Prediction X interview with Professor Ned Hall!