Christian Long

Hod Lipson: Builds “Self-Aware” Robots

In TED Talks on May 10, 2010 at 10:44 pm

Reflection by BENEDIKT K.

Original TED page w/ speaker bio, links, comments, etc:

Hod Lipson:  Builds “Self-Aware” Robots

For years robots have been proclaimed to be the future of housekeeping, our daily work and other small and extremely annoying to perform tasks. For about as many years however, the date for this revolution gets pushed further and further back, always being just out of reach of our current scope of things. Lipson’s talk very simply asks why this has been happening for so long, and what we can do to change it.

The main proposition is that we do not ourselves design these robots. Our own designs have been riddled with complications, from lacking ability to adapt to simply not working the way we want them to. So why not let the process that created life on earth as we know it, evolution, help us out in designing a better robot? If you put a whole bunch of motors, neurons and bars together, hopefully something will come out on the other end that is useful, and it does. Not that the robots in question are extremely advanced or even barely useful, they have been rewarded to move forward in this case, and they are amazing at moving forward.

Indispensable to this process is the process of the machine realizing its own basic form. Without it knowing how it looks, without some sense of self-awareness, it is extremely hard for these machines to begin running any sort of simulation on how to move forward. Even when the robot knows how it looks, it may not come up with the best ways of moving forward. But the interesting thing here is that the robot, which had 4 legs with knees and hips, did not come up with some form of a spider walk. It actually created a quite entertaining way of locomotion, which is very hard to describe, but it appears 3:55 into the movie if you are interested.

So if this robot comes up with a completely unexpected way of moving within a few minutes, or even if it took it a day, what else will they develop? At 5:25, a couple of cubes with motors inside them are shown replicating. The more raw material is provided, the more the robot can expand itself and multiply. So if nothing is provided but raw materials, with no reward for anything, the little buggers will simply multiply and grow.

So maybe we have been going at robots in the wrong way, by designing them ourselves. Evolution has created an extremely complex system that originated from a very simple set of conditions, while we have created fairly simple robots from very complex conditions and using highly advanced materials. It seems fairly logical that evolution should end up doing a better job at this task than we do. As long as we are in control of the situation, the rewards should be quite immense, and even if these robots were to “escape”, they wouldn’t exactly crawl over to the United Nations and proclaim themselves ruler over humanity.


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s

%d bloggers like this: