To end this series of blog posts, I’m going to talk about artificial intelligence and the need for engineers to broaden their knowledge on subjects outside their specific field of study.
The following TED Talk by Jeff Hawkins tells us exactly why we’ll never develop good artificial intelligence, if we’re to lazy to look at the underlying brain science.
Why A.I. needs brain science
Up until now, a lot of the current research effort has been focused on reproducing human behavior. Alan Turing famously launched the Turing test in his paper “Computing Machinery and Intelligence” (1950) in an attempt to provide a metric of performance for artificial intelligence.
In terms of this test, a machine is to be considered intelligent if it is able to imitate human behavior. In other words, a human being shouldn’t be able to distinguish that he/she is talking with a machine instead of another human.
Jeff Hawkins however, debates that true intelligence is achieved by having a memory systems that continuously stores what we experiences and plays back similar experiences to help us predict what will happen next.
Those who have read my latest blog posts shouldn’t be surprised that a lot of what our brain is doing is predicting the future. (Those who haven’t, I strongly recommend you read “Really, I’m just guessing! – Your brain” 😉 ).
Jeff Hawkins has a brilliant, yet simple way, to prove that we are really constantly trying to predict the future:
Because you’re intelligent and you speak English, you know what word is at the end of this …
sentence.
We don’t have to hear him say the last word. Out of experience we know what it will be, we hear it in our mind.
Why is this relevant?
The altered door experiment
[…] you have a door at home, and when you’re here, I’m changing it, I’ve got a guy back at your house right now, moving the door around, and they’re going to take your doorknob and move it over two inches. And when you go home tonight, you’re going to put your hand out there, and you’re going to reach for the doorknob and you’re going to notice it’s in the wrong spot, and you’ll go, whoa, something happened. It may take a second to figure out what it was, but something happened.Now I could change your doorknob in other ways. I can make it larger or smaller, I can change its brass to silver, I could make it a lever. I can change your door, put colors on; I can put windows in. I can change a thousand things about your door, and in the two seconds you take to open your door, you’re going to notice that something has changed.
Now, we all know the engineering approach: let’s make a huge database that contains every property of the door and then we’ll just check them one by one. But is that what’s really going on inside our brains? No!
What really happens is that we make a prediction every time we grab for that doorknob. We have an expectation of how the material will feel, what kind of shape it has, where we have to grasp, … If our prediction is of even slightly, we immediately know something has changed.
This is exactly why intelligence equals prediction and the domain of artificial intelligence needs some lessons in brain science.
Afterthoughts
This TED Talk was given in 2003, in the last decade Jeff Hawkins has made a lot of progress . For those of you who would like to learn more about this subject:
Video: Building Brains to Understand the World’s Data (Google Tech Talk by Jeff Hawkins)
His company specialized in machine intelligence: Numenta