TRYEngineering - AI Adventures - 12

ML is the Main Approach.
There's one other main type of machine learning called reinforcement
learning, which is used in robotics or video games. It teaches machines to
make decisions through trial and error in simulations. The machine gets points
for good actions, loses points for bad ones, and learns by trying to get the
highest score. Reinforcement Learning is similar to playing a new video game,
where you learn how you lose and win and use that to figure out how to get
better at playing it!
There are a few other ways that AI tends to learn, one of them being fuzzy
logic. This is when an AI learns by making decisions that aren't exact. It uses
rules and estimates to find the best solution. This is actually how washing
machines work! If your clothes are dirty, you wash them for longer; if they're
delicate, use a gentler wash cycle. You set your preferences, and fuzzy logic
considers all of them to figure out the best way to wash your clothes!
Another type is heuristic search. When an AI application uses heuristics,
shortcuts, or rules of thumb, it makes quick decisions and solves problems
faster. Instead of finding the best solution, it quickly finds one that's good
enough. You might use something like this when going to your friend's house or
somewhere else you know, picking the first path you think of instead of trying to
find the absolute fastest one.
There are many different ways to train AI that work better in different situations,
which gives people the freedom to pick the training method that best suits the
problem they're trying to solve!
TRYEngineering - AI Adventures

Table of Contents for the Digital Edition of TRYEngineering - AI Adventures
