Article: CMU online game will be used to help teach computers to see

CMU online game will be used to help teach computers to see. This is pretty cool. Basically, it's a game where one person tries to get another player across the internet to describe a picture, but the person describing doesn't get to see all of the picture at once.

The real point is to get enough pictures of objects at different angles with descriptions so that a computer can look at them all and learn what objects are by sight.

The only problem I see with this approach is that they'd need billions of images to make this worthwhile. For example, if you assume our eyes take in approximately 30 frames per second, at 60 seconds a minute, 60 minutes an hour, and 8 hours a day (average for a baby in the first year of their life), then in a single day, a baby sees 864,000 impressions of objects in the environment. Multiply that by 365 and you get 315 million impressions in the first year of life. So, you can imagine how many images are going to be necessary to teach a computer to do the same thing. I'd say, use video instead of static images.

Article: Bush wants alternatives to Darwinism taught in school

Bush wants alternatives to Darwinism taught in school -- for once I agree with something that President Bush says: "I think that part of education is to expose people to different schools of thought."

Profound, yet something that most schools don't do well. However, if you're going to open the schools up to different schools of thought, why stop at just adding Creationism to the curriculum? I see they're trying to call it "the" theory of Intelligent Design now, but what about other theories of intelligent design? Or maybe there are alternative evolutionary theories to Darwinism that we should discuss.

For example, what about aliens? I mean, couldn't aliens have created the universe as we know it? Isn't that "Intelligent Design"? Not exactly what Creationists are calling intelligent design, but worth considering. Shouldn't we tell our kids about that? How about a Human Endogenous Retrovirus (HERV) that punctuates evolution at times of stress in the population. This comes straight out of a series by Greg Bear, Darwin's Radio and Darwin's Children. There is even talk in that book about intelligence in the system, nature as a neural network of sorts.

Now, these ideas may seem crazy, but no more crazy than evolving from monkeys (where's the missing link?) or being thought up by some omniscient and omnipotent being, affectionately known as God to most of the western world. These ideas are just as crazy to the other group. My point is that I agree, we should teach many different viewpoints, and maybe even have debates in high schools.

Provoking independent thought in children is so very important, because then they won't be corporate drones in the future. And maybe they'll think so much they'll actually vote someone good into the presidency.

RoboMow Rocks!

A couple of months ago, my wife brought in a Sam's Club catalog and said to me, "There's something in here I thought you'd like," and smiled. I'm thinking, "Uh-oh, this is gonna cost us," knowing full well I'll probably love whatever she's going to show me.

So she opens the catalog and points to a little picture of an odd-looking lawn mower. Upon closer inspection, it appears that this lawn mower doesn't have a handle, and furthermore, it is sitting on what looks like a little docking station. I say, "No way!" and start to read more. It was, in fact, a picture of the Friendly Robotics RoboMow, an ingenious invention for mowing your lawn for you, and for about half the cost of a good riding lawn mower, which we were considering purchasing anyway. Not that our lawn really warranted one, of course.

The next week, RoboMow showed up at our house, and then the real work began. The way the system works is similar to an electric fence for dogs, with a perimeter wire to indicate its boundaries. Our lawn is a little larger than the 500ft of perimeter wire would allow, but we had anticipated this, having read a few reviews with pitfalls to avoid, so we had already purchased more wire and pegs to hold down the wire. It took us approximately 8 hours over the course of 2 days (we still needed more pegs, which a trip to Home Depot solved) to string the perimeter wire around the house and the lawn, in rain and sun (it poored and drizzled for a couple of those hours each day). But a few hours up front is still going to save us days over the course of having this little contraption.

After letting RoboMow charge for 24 hours, we went out to let it calibrate itself with the earth's magnetic poles for mowing in our yard. I'm guessing that it basically just has a compass that it uses to verify which direction it has gone and which direction it's going, to determine if it has covered all of your lawn. It uses a criss-cross pattern when mowing, hitting the grass at different angles so that after a few passes, it hasn't missed any spots. Then, once it thinks it is done, or the battery is too low, it follows the perimeterto the docking station to recharge.

We've run into a few problems with the system so far, and some can be avoided, while others are just based on the lay of the land. The book says to make sure that the lawn doesn't have any hills that are more than a 15% gradient. It can have trouble with those. We have a couple of steep spots, but nothing that RoboMow hasn't been able to handle yet. However, we have a water meter in the front that sticks up just a little bit, especially in dry weather, and a natural drainage channel in the back that stays a little wet, and RoboMow has gotten stuck in those areas a few times. We also have a swing-set area in the back yard with blocks around it that aren't 6 inches taller than the ground -- a requirement for the bumpers -- so sometimes it runs up on those, depending on the angle it comes in at. And finally, the dock really has to be level, or RoboMow will roll right out. That has been our biggest problem so far.

Overall, though, RoboMow does a great job on the lawn, and the coolness factor is definitely high. I think back to my days in Robotics class and remember how difficult it was to get our Lego robot to follow a line, much less give it blades and enough autonomy to mow the lawn! It's also good to see somebody reaping the rewards of robotic research. Hopefully it will give them some spending money to work on even better robotic tools.