I swear udacity is not paying me plug their courses, I am just an occasional user. Right now I am working through the Intro to Artificial Intelligence course and starting to dive into Bayesian statistics. This is an older course (2011), but it seems the principles still apply. The trend towards neural nets and deep learning still has probability at is core, which leads me to this question:
Is Probability at the core of all intelligence?
In the field of epistemology, many people defend inductive reasoning as the only path we have to knowledge. In a practical sense this is true, for example:
- the scientific method is essentially extremely rigorous inductive reasoning
- our day to day heuristics (the sun will rise tomorrow, my boss will ask me to come in this weekend [thanks Lumburgh])
- we ingest polls and surveys like gluttons to understand our social and political worlds
And yet, I have never been comfortable, epistemologically speaking, with deriving Knowledge (justified true belief if you will) from inductive reasoning.
This post was supposed to be about AI? You say. Now you're bringing up whether we know the sun will rise tomorrow as if you are at a bar full of undergraduates.
Fair, but here is my concern. I am a highly skeptical person when it comes to epistemology and metaphysics. Now here I am studying Artificial Intelligence when I do not have a firm footing on natural intelligence. I do not have an argument here, just a question.