Artificial Bee Intelligence
January 25, 2013
“Honey bees are fascinating creatures,” writes Ben Coxworth. “They live harmoniously in large communities, divided into different castes, with some of the worker bees heading out on daily expeditions to gather nectar and pollen from flowers.” [“Scientists hope to put artificial bee brains in flying robots,” Gizmag, 2 October 2012] Those daily pollen-gathering expeditions have fascinated scientists for years. Bees have evolved the ability to make those trips to and from the hive as efficient as possible and, as Coxworth notes, “A study has suggested that the efficient method in which bees visit those flowers could inspire the improvement of human endeavors such as the building of faster computer networks.” As a result, there has been a spate of news stories about how scientists are trying to recreate some of the bee’s most beneficial talents. Coxworth describes one such effort. He writes:
“Scientists from the Universities of Sheffield and Sussex hope to build a computer model of the honey bee’s brain, with the ultimate hope of using it to control tiny autonomous flying robots. The project is called Green Brain – a tip of the hat to IBM’s Blue Brain Project, the aim of which is to create a computer model of the human brain. The Green Brain team, however, aren’t actually trying to recreate all of a bee’s mental processes. Instead, they’re focusing on the systems that control its vision and sense of smell. Also, unlike the Blue Brain scientists, they’re not using supercomputers to create their model. In order to get the performance they’ll need out of desktop PCs, they are using high-performance GPU (graphics processing unit) accelerators.”
Coxworth asks a good question, “Why would anyone want a bee-brained flying robot?” It turns out there are, in fact, good reasons. He explains:
“In the same way that honey bees can sniff out and visually identify flowers, it is hoped that the autonomous robots could be used to trace odors or gases to their sources. Not only could this have applications in fields such as environmental monitoring, but it could also prove useful for things like search-and-rescue operations.”
Coxworth also notes that if bee colonies continue to collapse, artificial bees may be needed to pollinate crops. Coxworth’s Gizmag colleague, Dana Borghino, reports that researchers at Queen Mary University of London are also studying bees and how they can “fly from flower to flower and then come back to their hives expending the least amount of time and energy.” They hope to apply their bee knowledge to the well-known “traveling salesman” problem. [“How bumblebees might get you faster overnight deliveries,” 24 September 2012] Borghino notes that the traveling salesman problem is “employed by all courier companies in day-to-day operations to deliver and pick up goods using the least amount of time and fuel while still visiting all the required destinations; electronics manufacturers use it to create better and more powerful microchips; and geneticists use it to sequence DNA.” He continues:
“Computationally, this problem is as hard as they come. The most direct, ‘brute force’ approach to finding a solution would be to try out every possible combination but, in practice, this is unfeasible: with just 20 cities on the itinerary, there would already be 60.8 quadrillion comparisons to carry out – impractical even for today’s computers. … If we want to do better, we need to look elsewhere. … Whenever researchers are stumped by a problem, looking at nature’s solutions is often a good option. … Now, judging from the recently announced findings, it looks like individual bees might also be able to contribute something to the field. A group of researchers from the Queen Mary, University of London has recently shown that the common bumblebee can solve a problem analogous to the traveling salesman’s by using a simple iterative approach that requires no real number-crunching – only a tiny bee’s brain.”
It turns out that bees use a combination of trial and error and selective memory to discover and remember the shortest path between hive and flower. Borghino writes, “The process is remarkably similar to the widely adopted ant colony optimization algorithms, in which the increased probability of updating the shortest route is substituted by a stronger pheromone trail.” He concludes that when bee behavior is modeled appropriately it “could result in an algorithm that is more generalized, much more flexible and well-suited to situations in which the number of resources, their spatial configuration and their reward values are changing over time.”
Peter Rothman reports that “a full connectome map of the honey bee brain has been developed.” As a result, “Artificial insect minds may be the first true Artificial General Intelligences (AGIs) available for commercial use.” [“The Buzz of the Hive Mind: Artificial Bee Brains,” hplusmagazine.com, 2 October 2012] Rothman continues:
“Researchers at the Institut fur Biologie-Neurobiologie Berlin, Mercury Computer Systems GmbH, SRI International, and the Zuse-Institut Berlin have developed a three-dimensional average-shape atlas of the Honeybee Brain. This atlas could allow the development of a full robotic insect capable of imitating the full social capabilities of an insect operating in its natural environment. This would open the door to entirely new ways to study insects using a robotic insect to enter and participate in activity in the hive while transmitting information back to human scientists in real-time. Robotic insects could also be used to help endangered biological insect colonies more efficiently locate food sources or offer enhanced protection against predators, disease, etc., for example to rescue specific bee colonies in danger of collapse. An important point to understand here is that the honey bee brain is a general purpose bio-information processor capable of complex spatial perception and recognition of visual and olfactory stimuli in open environments. The bee brain can perform real world pattern recognition and categorization in the presence of noise and interfering or competing signals, it provides goal based guidance and control of a winged flying body, and bees operate in groups demonstrating group goal based behaviors such as directed and optimal search by using social communication and ‘swarm’ organization. … The widespread availability of bee based AGI will allow these problems to be solved using this same approach providing better solutions for numerous problems in routing, path planning, and resource allocation.”
Harvard University researchers have been working on a robotic bee for the past five years. Evan Ackerman reports:
“Five years is a long time in the fast-paced world of robotics, but when you’re trying to design a controllab
le flying robot that weighs less than one tenth of one gram from scratch, getting it to work properly is a process that often has to wait for technology to catch up to the concept. The RoboBee has been able to take off under its own power for years, but roboticists have only just figured out how to get it to both take off and go where they want it to. Or at least, they’re getting very, very close.” [“Harvard RoboBees Learn to Steer, Mostly,” ieee spectrum, 8 October 2012]
Obviously, there is more than novelty and curiosity spurring the development of robotic bees powered by artificial brains. Anytime, however, that someone raises the possibility of releasing masses of autonomous robots into the real world, someone is going raise the spectre of catastrophe. Perhaps the next big disaster film will be “The Swarm II: Robot Bees.” If you’ve forgotten the move entitle “The Swarm” you’re probably not alone. The following trailer gives you a pretty good indication of why it was forgettable.
Personally, I’m not predicting that a swarm is coming or that artificial insects will represent a threat to mankind. I’ll let others, like researchers at the Centre for the Study of Existential Risk (CSER), worry about such things (see my post entitled Artificial Intelligence: Is there Peril in Deep Learning?). If teams creating insect-size robots and insect AGI cooperate, don’t be surprised if you one day see an interesting looking bug coming to save your life during a natural disaster.