Artificial intelligence is a pillar of technological research and progress. Tech innovation depends on machines who can think critically and develop novel solutions to problems. AI (and its close relative, machine learning) is so pervasive now that it no longer exists only in the tech fields. In recent years, AI has become increasingly important in the legal field as well. But where did it start?
There are many ways to trace the history of artificial intelligence. One of the most important developments in the concept’s early history came from Thomas Bayes. Bayes was an English mathematician and minister in the mid-18th century. He was not a very prolific author, and his works were only published after his death. Little else is known about Bayes. His work, however, demonstrated a fascinating new idea in probability.
It seems almost poetic that little is known about Bayes’ life other than his work. AI is slowly displacing the “cult of the expert”, among other things. It seems fitting that a man who contributed to the ideas behind AI is rendered somewhat nondescript by history. But what were those ideas? Let’s take a look.
Bayes’ famous theorem is a type of probability calculation. It uses prior knowledge of conditions related to an event to find the probability of that event occurring in the future. You might say it was the first time mathematics was used to shore up the validity of the phrase “educated guess”. The equation developed by Bayes and his colleague Richard Price allows us to use evidence to update inferences. In other words, once a probability estimate is made, new information can be incorporated into later calculations of that same probability. Sound familiar?
For a visual aid, here is a breakdown of the equation Bayes and Price developed:
P(A) and P(B) are the probabilities of observing A or B independent of the other
P(A|B) is the probability of observing A given that B is true
P(B|A) is the probability of observing B given that A is true
The theorem seems like a basic concept. Like the pioneering work of Pascal, Fermat, and others, this theorem seeks patterns in the world that may allow us to predict the future. We will soon see, however, that some of the calculations made possible by Bayes’ theorem are not as intuitive as one might at first think.