How is Bayes' Theorem used in artificial intelligence and machine learning? As an high school student I will be writing an essay about it, and I want to be able to explain Bayes' Theorem, its general use, and how it is used in AI or ML.
Bayes theorem states the probability of some event B occurring provided the prior knowledge of another event(s) A, given that B is dependent on event A (even partially).
A real-world application example will be weather forecasting. Naive Bayes is a powerful algorithm for predictive modelling weather forecast. The temperature of a place is dependent on the pressure at that place, percentage of the humidity, speed and direction of the wind, previous records on temperature, turbulence on different atmospheric layers, and many other things. So when you have certain kind of data, you process them certain kind of algorithms to predict one particular result (or the future). The algorithms employed rely heavily on Bayesian network and the theorem.
The given paragraph is introduction to Bayesian networks, given in the book, Artificial Intelligence – A Modern Approach:
Bayesian network formalism was invented to allow efficient representation of, and rigorous reasoning with, uncertain knowledge. This approach largely overcomes many problems of the probabilistic reasoning systems to the 1960s and 70s; it now dominates AI research on uncertain reasoning and expert systems. The approach allows for learning from experience, and it combines the best of classical AI and neural nets.
There are many other applications, especially in medical science. Like predicting a particular disease based on the symptoms and physical condition of the patient. There are many algorithms currently in use that are based on this theorem, like binary and multi-class classifier, for example, email spam filters. There are many things in this topic.I have added some links below that might help, and let me know if you need any kind of other help.
It helps in improving the efficiency in solving real world problems. When Air France flight disappeared in Atlantic Ocean in 2009 then Scientists developed a Bayesian model to predict the location of the plane. The model took in factors such as the expected flight plan, weather, ocean currents and other external factors. The model then mapped a probability to a 50 mile radius around the expected crash zone. Each point within the 50 mile circle was assigned a probability of the plane being located there. The he model used a large data set of information that was updated continuously as the search team entered results everyday after search a specific location. Within days of implementing this model, the plane was found. This shows how statistical models and theory can help improve efficiency in solving real world problems. Link for article
Since you are a highschool student I will try to express it easier. It is a problem for a machine to make a decision if you haven't given that information to it before. You should think of every cases while programming. But sometimes there can be so many cases, here Data Mining, Neural Networks, Fuzzy Logic etc are used withing AI. It saves your time and system is learning itself with enough examples given at the beginning and deciding itself.
Here in this link you can find an article about Bayesian learning. Example on p.33 is what you need I guess.