What Tech Can Learn from the Fruit Fly's Search Algorithm - Facts So Romantic

Ask, and it shall be given you; seek, and ye shall find; knock, and it shall be opened unto you." Verse 7:7 from the Gospel of Matthew is generally considered to be a comment on prayer, but it could just as well be about the power of search. Search has become one of the key technologies of the information age, powering industry behemoths and helping us with our daily chores. But that's not where it ends. Scientists are starting to understand that search powers much of the natural world, too. Saket Navlakha, of the Salk Institute for Biological Studies, works at the "interface of theoretical computer science, machine learning, and systems biology," a field, he told me, that he and his colleagues are calling "algorithms in nature." Evolution needs algorithms just as software engineers do, Navlakha says, because it "has also had to deal with building efficient, reliable, low-cost systems that help animals and organisms survive." His hope is to find in nature "new ideas and new engineering principles" that can be exploited by human scientists and engineers. In a study published on Friday, Navlakha and a couple colleagues, Sanjoy Dasgupta and Charles F. Stevens, did just that. They found that the fruit fly brain had some valuable lessons for anyone developing similarity search algorithms. Stevens had been studying fly neural circuits, specifically how they associate different behaviors, like approach or avoidance, with odors in the environment. "When he started telling me about it," Navlakha says, "I realized that what the fly needs to do is do something like a similarity search.