Machine Learning

Machine Learning (ML) is set of few basic ideas. Unlike the traditional algorithm, ML has the non linear chops to model most phenomena we see around us.  ML is about what we want, the results of our actions, how to achieve our goals and how the world will change.  Its a  paradigm shift.

There are different schools of thought in ML that help us answer these questions: How do we learn, is there a better way, can we trust what we learnt, what can we predict. The significant 5 schools/themes are:

1. Symbolists: inverse of deduction
underlying algorithm: inverse deduction

2. Connectionists: reverse engineer the brain
underlying algorithm: back propagation

3. Evolutionaries: simulate evolution on the computer
underlying algorithm: genetic programming

4. Bayesian: learning is a form of probabilistic  inference. statistics
underlying algorithm: bayesian inference

5. Analogizers: extrapolating similarity judgement.
underlying algorithm: support vector machines

The master algorithm is a combination the key features of all these 5 algorithms that would derive all the knowledge in the world (a unified theory that makes sense of everything we know today) and solve the hardest problems that plague humankind today.

Master algorithm is to Machine Learning what the Standard Model is to Particle Physics.

Advertisement