Increased Shipping to Accelerate Climate Change as Arctic Warms
Image of the Day: The Dragon's Eye

Bees Solve Complex Problems Faster Than Supercomputers

Bees_experiments_flowersIn a new study, researchers report that bumblebees were able to figure out the most efficient routes among several computer-controlled "flowers," quickly solving a complex problem that even stumps supercomputers. We already know bees are pretty good at facial recognition, and researchers have shown they can also be effective air-quality monitors. 

Bumblebees can solve the classic "traveling salesman" problem, which keeps supercomputers busy for days. They learn to fly the shortest possible route between flowers even if they find the flowers in a different order, according to a new British study.

The traveling salesman problem is a  problem in computer science; it involves finding the shortest possible route between cities, visiting each city only once. Bees are the first animals to figure this out, according to Queen Mary University of London researchers.

Bees need lots of energy to fly, so they seek the most efficient route among networks of hundreds of flowers using angles of sunlight, which helps them find their way home, researchers say. To do this, their tiny brains must pack a powerful memory.

To test bee problem-solving, researchers Lars Chittka and Mathieu Lihoreau tested bees’ response to computer-controlled artificial flowers. They wanted to see whether the bees would go after the flowers in the order in which they were discovered, or if they would figure out the shortest route among all the flowers even as new ones were added. The bees explored the locations of the flowers and quickly figured out the shortest paths among them, according to a Queen Mary news release.

This is no small feat, especially considering the tiny size of bee brains. When it comes to certain types of intelligence, size apparently does not matter.

Earlier this year, researchers showed that bees recognize individual faces because they can make out the relative patterns that make up a face. The new research further suggests bees are highly sophisticated problem solvers, and that better understanding of their brains could improve our understanding of network problems like traffic flows, supply chains and epidemiology.

The research will be published this week in the journal The American Naturalist.

Casey Kazan via Queen Mary University of London



This makes me wonder if bees might not have some sort of "distributed computing" property, somehow sharing complex processing tasks throughout an entire hive. With so much still unknown about how bees' senses and other systems work, it's hard to tell. (If they have electrosense, I've never heard about it, and I have a more than passing interest in that sense.)

It's either that, or their individual brains are hardwired to figure out these kinds of complex problems. While certainly possible, that seems even more bizarre than the distributed computing theory.

Still, I don't think we'll be seeing bees helping TV programming executives select the fall schedule any time soon.

Animals & insects are much smarter than we give them credit for.

Bob: Are you implying that bees have a psychic hivemind? Because these bees are doing this independently, not as a group.


JJJ: Kind of. Not so much "psychic," as I'm not convinced that bee consciousness is that powerful (I'm not even convinced that human consciousness is that powerful), but rather closer to the mode of a wireless computer network. There are fish that communicate by electrosense; the hypothesis I'm suggesting is that bees might do the same. (Emphasis on "might.")

From the article:

In a new study, researchers report that bumblebees were able to figure out the most efficient routes among several computer-controlled "flowers," quickly solving a complex problem that even stumps supercomputers.

I call bullshit on this. While the brute force approach to the "Traveling Salesman" problem is an O(n!) problem, meaning that the computational load increases as the factorial of the number of flowers (or cities), algorithms for exact solutions exist which can handle many 10s of thousands of flowers (cities) very quickly. And approximate solutions are much faster than that. i.e. the TSP does not "confound" supercomputers unless the algorithms they are running are extremely naive.

I'm sure that the bees' performance is impressive considering their tiny brains. But let's not go off the deep end with our claims.

How many flowers? Did the bees find the exact solutions, or just approximate ones?

Bit by bit we will learn to value nature and its myriads of solutions to survival, we are not special, we are just another variation, better at some things, worse at others.

The question is, will we learn to value nature while there's still any nature left to value?

If this must be seen as a competition, I think artificial intelligence is not ready to go head to head with nature's intelligence. Spiderwebs anyone?

Once again, Asimov (among many others) is proven right. Naturally occurring brains are, quite literally, molecular computers. Complex chains of reactions are encoded for various responses and combined in millions of different ways in the blink of an eye (sometimes literally) to carry out simple mechanical functions with the maximum degree of control and efficiency. The greatest supercomputer on earth is dumber than the lowly insect, because at least the insect can improvise when things do not go perfectly according to plan. Computers just pop up an 'error report' and wait for you to fix the problem for them.

Once again, Asimov (among many others) is proven right.

Isaac Asimov. My long lost friend. He never knew me from Adam. We never corresponded. Yet when he died in 1992, I felt a sense of loss which stays with me to this day.

But I disagree with your assertion that computers, in general, "just pop up an error report". Your desktop box might do so. But certainly, embedded systems do what makes sense. If sensory input looks to have gone haywire, or results don't pass minimal sanity tests, they switch to a reasonable failure mode. It might involve cutting the engine, if the situation warrants (high coolant temp). But more likely, it will involve putting the engine into an open-loop mode, which works less efficiently than a carburetor might, but still passably well.

Computers can ad-lib when things do not go according to plan. Rather better than bees do, I would imagine.

Any intelligence which might emerge from our current technology is likely to look quite *different* than that of, say, bees, or of primates. But such intelligence should be considered no less noteworthy.


A small jumping spider hunts insects many times larger than itself using complex hunting strategies but has a brain smaller than a grain of salt.

A small jumping spider hunts insects many times larger than itself using complex hunting strategies but has a brain smaller than a grain of salt.

A guy I know from the bar makes plenty of conquests with a brain lots smaller than that. Then again, I guess not everything about him is small. I wouldn't know, you understand. But I'm just say'n.


In the early days of ai research a group of students beat the best computer available to the shortest route between US state capitals with nothing simpler than intuition and a ball of string.
It is simply arrogant to assume that in the fifty odd years of electronic computation that we could exceed three billion years worth of non linear factorial organic computation, evolution.
Evolution doesn't make assumptions, of worth, get sidetracked or discount what appears unfeasible, it simply works. Evolution is of itself the best method of reaching the best result, or evolution as we know it has surpassed all other evolutionary mechanisms.
Evolution evolved to be the best.

"Computers just pop up an 'error report' and wait for you to fix the problem for them."

You have to realize that is not at all an accurate description of computer science.

For one, most algorithmic problems do not require much ad libbing, and simply necessitates fast (in terms of computational complexity) series of decisions to reach an accurate solution. The traveling salesman problem, for instance, is interesting not as a test of the computational velocity of super computers, but rather serves as a challenge to the programmer to concoct a reliable heuristic algorithm to compute MOST variations of inputs in the fastest time possible. (The problem is proven NP-hard I believe, so it is impossible to find a solution to solve all cases in polynomial time). Running the proper algorithm, a supercomputer can easily outperform a bee. If this article's facts are proven reliable, then its conclusion is not. The bee's ability would be impressive in the sense that it can create a far more intelligent solution to the problem than we human programmers have developed, and has nothing to do with the speed of our best computers.

(The slowest computer in the world running the right algorithm with a large enough input size will outperform the fastest super computer running a naive solution.)

And as for the computer's inability to adlib? That's just ridiculous. A computer is only as useful as the programs we run on it, and we very well can run this:


Have you considered whether bees might be using self-made chemicals that only their hive-mates recognize in the same way that ants do? This seems more likely to me than an electrosense given that bees are more-closely related to ants than to fish. Just a thought...

The collection of insects, in this case a hive of bees, have been in existence almost unchanged for millions of years. Nature will have given them essential and sophisticated tools for survival. Just because we don't know what it is or understand it when we see it doesn't mean that it doesn't exist. Eg we know that bees talk to each other by a complicated "dance".

Dolphins have a language and names yet we persist in treating them as stupid animals. All this does is show our ignorance.

Bees tell each other where the best flowers are in a pre defined dance consisting of body movements, Question here is are the bees in the experiment alone or part of a hive, we know that they have a limited neural capacity do the solution to the quick travelling salesman problem can be an emergent result of the dance pattern, I wish I was a biologist to work on the subject unfortunately I'm a computer scientist.

Title is misleading and I am very sceptical about what article claims. Do bees see thousands of flowers then decide where to start? I don't think so. What if wind moves the flowers after the bee begins? Bad science.

Who designed the bees brain in such way that to solve complex problems ? Is it evolved automatically ? What will be the mathematical probability to evolve this highly sophisticated problem solving capability as in the evolution theory?

Yes there must be a designer behind this... Try to find out the designer - the GOD - almighty... and try to find out the meaning of life...

And thy Lord taught the bee to build its cells in hills, on trees and in (men's) habitations..... there issues from within their bodies a drink of varying colours, wherein is healing for mankind. Verily in this is a Sign for those who give thought'.
(Translation of Quran 16:68-69)

This is interesting but not too surprising either. This is not about intelligence per se but about evolution at work. As the author points out it would be very energy intensive to fly willy-nilly. Over the untold millennia that they have been evolving a method of intuitively seeing the best routes would have been strongly selected for. Those bees that did this routing better produced more off-spring and thus would become the dominant species.

I wish people would stop using the term "tiny brain" people seem to forget that bees are a hell of alot smaller than humans, so of course they would have smaller brains. But when we take into account the size of the animal the brain isn't all that small. The size of the brain is perfect for the size of the animal. Also the size of a brain has no effect on it's ability.

"Computers just pop up an 'error report'"
Well actually no they don't. They either pop up an error message and wait for a technician to address it or they ignore the problem and continue on as they would except without whatever caused the error.

Which of these a computer does was up to the earliest developers. To have it ignore problems is very convenient for non-technicians just trying pursuing leisure activities while halting at slight errors is very convenient for the person actually doing the programming, because they can then tell exactly what kind of error they made. If you're working with something really big and complicated you probably want all the help you can get identifying quite where something goes wrong or else you're likely to miss it altogether.

So the ones that stop right away at problems really just force the programmer to do a better job, and get something that actually does what it is supposed to for it. And really, when you've got the possibility of calling someone up if you run into some obscure problem then you should have the machine halt at errors- that way at the very least the developers might ever hear about it instead of the thing just chugging along recording the wrong money values at a bank or whatnot. If instead you've got something you're sending into space- well, maybe you want that to only even send a short transmission about which error it ran into so you are aware of it, while the thing keeps doing whatever task it is supposed to.

Too bad they're all going to die off before we figure out how they're doing it.

Since bees can communicate to each other where stuff is and navigate via the sun's angle/time of day, and magnetism of the earth. Maybe they some how include that in their calculations?

It could be what we're missing.

Verify your Comment

Previewing your Comment

This is only a preview. Your comment has not yet been posted.

Your comment could not be posted. Error type:
Your comment has been posted. Post another comment

The letters and numbers you entered did not match the image. Please try again.

As a final step before posting your comment, enter the letters and numbers you see in the image below. This prevents automated programs from posting comments.

Having trouble reading this image? View an alternate.


Post a comment

Your Information

(Name is required. Email address will not be displayed with the comment.)