This is a teaser for Act III and Act IV
Remember, all machine learning models are just algorithms.
The difference is that instead of being explicitly programmed, they adapt specifically to the data, making them able to solve much more abstract, unstructured and unpredictable problems.
But there is a whole universe of algorithms that are used to solve problems just as interesting.
Within machine learning
Beautiful Deep Learning, which you will deep dive into soon in Act III.
Beyond machine learning
⚘ Classical algorithms: the bread and butter of computer science. All of these algorithms form the computational backbone for machine learning algorithms.
⚘⚘ Heuristic algorithms, which optimise problems by imitating the principles of evolution. If you like deep learning, you will love heuristic algorithms.
⚘⚘⚘ Probabilistic algorithms exploit the mechanics of probability to solve problems that are simply too hard for traditional hard coded algorithms. Read more here.
ML algorithms learn to approximate the answer by essentially learning from the answers.
Deterministic algorithms are meticulously written to tackle … through instruction based, programmed logic. Deterministic algorithms KNOW THE EXACT process to get the answer to any given problem. My favourite classic example to start with:
Not only do deterministic algorithms give the same answer every time, but they give the true, underlying, correct answer.
I.e: shortest job first