The conversation of earthquake modeling reminded me in many ways of our discussion for modeling alien life and extraterrestrial contact. Meade comments on the complexity of the polycrystalline aggregates, for example, which comprise just one aspect of the uncertainty and difficulty entailed in predicting earthquakes. This, like our data collection on the cosmic microwave background, seems to be a trove of rich data information but may be too complex for our modeling and understanding to fully interpret intelligibly. Meade comments optimistically on the usage of deep learning and AI to modify the predictive models levied to forecast earthquakes which also had proliferative impacts that benefited other infrastructural sectors of society. I was wondering how trying to solve the problem of alien life might benefit us in other parts of society that we might not foresee right now.
I had a question for professor Murphy in her endeavor to apply statistics to bolster mobile health insights. The concept of hypothesis-free science seems to fly in the face of some probabilistic modes of thinking but immensely useful for approaching conundrums with several unknowns unknowns. I was wondering to what extent can we aggregate data to recognize features in useful ways? Again how do we make the data that we encounter using these deep learning models useful to the people that didn't develop them or simply the most amount of people period?