Monte Carlo methods are ones which exploit the idea that if you randomly try something enough, eventually you’ll converge to the true probability. For example, if you flip a coin 10 times, you might get heads 70% of the time, but as you keep flipping the coin, you should get closer to a 50/50 split.1
What can you do with this stupid brute force approach?
- Monte Carlo Tree Search to solve games (this was used in AlphaZero to beat the top Go player)
- Estimate the value of \pi
- Simulate gnarly distribution
If you have more ways that MC has been applied, let me know!