Definition
- Broader notion of computational artifact that learn based on experience
Supervised Learning: Based on the input and labeled output, predict the future output given the future input
Induction and Deduction:
- Deduction is based on the notion that we already know the rules, the formula. Most AI used pre-definitive rules/formula to get artifical inteligence
- Induction is from the existing input and output, we want to guess what the rules is. And the rules we define, is not definite. But generalized rules(induction).
Unsupervised learning: We want to take a concise descriptive compact to structured all inputs. Structured based anything. Given 100 people, structured based on gender, based on beard, based on additional hair, etc.
- Reinforcement Learning: The machine told what's the outcome. It doesn't know what to do with each step, good or bad (as opposed to supervised learning). The machine then have to figured it out, what's steps to do along the way to get to win (in case of a game), or in general overall result.
-