Slashdot Mirror
Smart Bees Continue to Draw Interest
Roland Piquepaille writes "Researchers from University College London (UCL) have shown that bees are able to solve complicated color puzzles. In their study, bees were trained to find artificial flowers containing a nectar reward and colored in blue. Then they removed the nectar and put at random artificial flowers illuminated by yellow, blue, yellow and green lights. But wherever the blue flowers were, the bees continued to select them even if there was no longer a reward. These findings may soon lead to the design of sophisticated visual systems for autonomous robots."
The story here isn't that bees are smart - this is simple pavlovian association going on, not any sort of problem solving. The story is that they are able to recognize the color of objects that are complexly lit.
They keep going for it after it stops paying? Funny, we hear that octopus are smart because they don't.
Someone confused "memory" with "intelligence" again.
You can't take the sky from me...
In other news my dog can find and separate his medical pills even when they're complexly hidden in cheese. He has self-trained to reject the pill and spit it on the floor while still consuming the cheese.
This may possibly soon lead to complex robotic cheese/pill separators for robots that have colds.
1) Imagine a beowulf cluster of apiaries with such bees.
2) a) Train bee
b) ?????
c) Profit!
3) I for one, welcome our new smart color bee overlords
4) In Soviet Russia, the bee sees *you*
5) I'm allergic to bees, you insensitive clod!
6) Trained bees? Won't somebody think of the children!
7) Netcraft confirms: the bees are dying
8) All your nectar are belong to bees.
9) Those are right-wing, neocon bees.
10) Bees teh sux! Wasps 4evar!
11) All I wanted was some frickin color-distinguishing bees with some frickin laserbeams!
12) In my day, we had bees that could distinguish flowers by taste or smell, at 100 yards, in a snowstorm. And liked it!
13) I don't see any b33#$(*&#$(*#& NO CARRIER
Now that they have found a way to get a bee attracted to something, maybe they should look for a way to repel bees from things like people and picnics!
DEAD DEAD DEAD DELETE ME
5 Comments in 45 minutes? This must be some strange use of the word "Interest" that I was previously unaware of.
And coming up next, can bees think? A new study confirms that, no, no they cannot.
... bees have 'eyes' and are capable of 'sight'.
The Dutch will inherit the earth. If not, we'll settle for a bit of ocean. Beta delenda est!
I personally don't see what all the buzz is about...
The view was horrible and the smell was even worse; Julie severely regretted becoming a proctologist.
14) Oh, yeah, what are you gonna do? Release the smart dogs? Or the smart bees? Or the smart dogs with smart bees in their mouth and when they bark, they shoot smart bees at you?
... or, in fact, that "Bees solve complex colour puzzles." I think most readers would assume, when they see a title like that, that there would be some sort of, you know, puzzle involved in the experiment. I'll give them complexity; this is not a trivial vision task. "Puzzle" suggests, to me, something more like maybe learning to choose the middle wavelength of a set of three colors, at the very least. Maybe even logic or something.
I know bees can find flowers in the forest. I see them when I go hiking. The breakthrough here is finding a way to approximate the same sort of stimulus in a laboratory setting. I don't know very much about insect neuroscience in particular, but data from the study of higher animals (monkeys, humans, rats, cats, some birds, and others) strongly suggests that this type of statistical pattern extraction is what is taking place. Computer modeling systems involving neurally-inspired architectures (neural nets) also indicate that this sort of behavior arises naturally from these systems and so is a computationally plausible explanation as well. I'm very surprised to read that they didn't think this was a possible model for bee brains as well. We may some day get as far as being able to "construct useful behavior," but all this study can do is shed some insight into how a visual system can construct useful representations from ambiguous sensory information. The construction of useful behavior in this study happened when they trained the bee to go prefer a particular color. Then they used that useful behavior to find out just what the bee could see.In conclusion: a very useful study, that gives us important new information to help us figure out how visual systems in particular, and bee brains in particular, extract useful information from confusing data. But all of the grandstanding about complex puzzles and computer vision? I say they're blowing steam out their... um... stingers.