How does artificial intelligence learn? - Briana Brownell
711,270 Views
16,410 Questions Answered
TEDEd Animation
Let’s Begin…
Today, artificial intelligence helps doctors diagnose patients, pilots fly commercial aircraft, and city planners predict traffic. These AIs are often self-taught, working off a simple set of instructions to create a unique array of rules and strategies. So how exactly does a machine learn? Briana Brownell digs into the three basic ways machines investigate, negotiate, and communicate.
Additional Resources for you to Explore
The history of machine learning is full of fascinating machines and programs that use the three major techniques we have covered in this video.
In fact, the term “machine learning” comes from a checkers-playing AI created by Arthur Samuel. Although checkers is much simpler than games like Chess and Go, which have both been more recently tackled by machine learning methods, it still has 500 billion billion – that’s 5 with 20 zeroes – possible positions. This complexity made it impossible to solve using exact methods. Instead, Samuel used a formula with six variables that were related to winning the game, and tracked them as his program played itself. His checkers-playing AI could train itself to pay competently in 8-10 hours – a huge feat for a computer in the 1950’s. You can read more about his program here. Since then, machine learning researchers have tackled many other games including: Go, Poker, and Atari console games.
But machine learning isn’t all about games. Many of the techniques that researchers use to solve these kinds of problems can be applied to many other areas, including self-driving cars, diagnostics, and language translation.
In fact, the term “machine learning” comes from a checkers-playing AI created by Arthur Samuel. Although checkers is much simpler than games like Chess and Go, which have both been more recently tackled by machine learning methods, it still has 500 billion billion – that’s 5 with 20 zeroes – possible positions. This complexity made it impossible to solve using exact methods. Instead, Samuel used a formula with six variables that were related to winning the game, and tracked them as his program played itself. His checkers-playing AI could train itself to pay competently in 8-10 hours – a huge feat for a computer in the 1950’s. You can read more about his program here. Since then, machine learning researchers have tackled many other games including: Go, Poker, and Atari console games.
But machine learning isn’t all about games. Many of the techniques that researchers use to solve these kinds of problems can be applied to many other areas, including self-driving cars, diagnostics, and language translation.
About TED-Ed Animations
TED-Ed Animations feature the words and ideas of educators brought to life by professional animators. Are you an educator or animator interested in creating a TED-Ed Animation? Nominate yourself here »
Meet The Creators
- Educator Briana Brownell
- Director Champ Panupong Techawongthawon
- Narrator Safia Elhillo
- Music Ambrose Yu
- Sound Designer Ambrose Yu
- Director of Production Gerta Xhelo
- Editorial Director Alex Rosenthal
- Producer Bethany Cutmore-Scott
- Editorial Producer Dan Kwartler
- Script Editor Eleanor Nelsen , Alex Gendler
- Fact-Checker Eden Girma
