Member-only story
Basics to Deep Learning: Teaching Computers to Understand the World… Like Us
Imagine teaching your grandma how to use personal assistants. You wouldn’t hand her a complex manual, right? Instead, you probably would show her examples, answer her questions and let her experiment. That is exactly how deep learning works!
Remember those sci-fi movies where robots ruled the world?
Early 90s was an exciting time for AI because it was the first time a computer “Deep Blue” was able to defeat the chess world champion. Deep Blue is an example of an expert system meant to mimic a human subject matter expert by hard coding many rules and took hundreds of thousands of engineers to make it, which became its main limitation.
Some tasks that are simple for people, it was very difficult to write rules (programs) that a computer could follow. As an example, see the picture below and think of words that describes it and the humor behind it.
For an adult with exposure with this popular culture probably it isn’t difficult but how do you teach this to someone who speaks a different language, or a new born and let alone an artificial intelligence.
A popular learning technique is trial & error, where people are exposed to lots of examples, then make predictions on what the correct answer is, and if they are wrong…
