Are bug brains the key to AI?
- By Susan Miller
- Jan 07, 2019
Current technologies used for artificial intelligence tend to be large, heavy and power hungry – exactly the characteristics that make deploying AI to the field problematic. One solution might be found in insect brains.
The "impressive computational capabilities" of very small flying insects comes in a tiny package – sometimes of only a few hundred neurons, the Defense Advanced Research Projects Agency said in an invitation for new research concepts called Microscale Bio-mimetic Robust Artificial Intelligence Networks (µBRAIN). Insects also seem to be able to have "subjective experiences," implying they can infer, generalize or abstract information from their experiences to solve problems -- the fundamental processes underlying AI.
Insights gained from research into insect brains could lead to new ways of solving AI-based problems, DAPPA hopes, as well as new computing paradigms that will improve AI by considerably reducing training times and power consumption.
The primary goal of µBRAIN, the agency said, is to understand the computational principles, architecture, physical interactions and mechanisms of small biosystems that operate under extreme size, weight and power-consumption restrictions. Once researchers understand the miniature but highly integrated sensory and nervous systems in insects, DARPA hopes they can develop computational models and appropriate hardware that mimic those refined functions.
The research agency is not interested in learning how insect brains work but rather in discovering generic and generalizable computational models inspired by these insects that will provide insights for new efficient approaches to contextual AI. DARPA expects that achieving acceptable performance with low size, weight and power requirements will demand "a different set of guiding principles" across physics, chemistry, components, architecture and computational models.
Proposals are due Feb. 4. More information is online here.
Susan Miller is executive editor at GCN.
Over a career spent in tech media, Miller has worked in editorial, print production and online, starting on the copy desk at IDG’s ComputerWorld, moving to print production for Federal Computer Week and later helping launch websites and email newsletter delivery for FCW. After a turn at Virginia’s Center for Innovative Technology, where she worked to promote technology-based economic development, she rejoined what was to become 1105 Media in 2004, eventually managing content and production for all the company's government-focused websites. Miller shifted back to editorial in 2012, when she began working with GCN.
Miller has a BA and MA from West Chester University and did Ph.D. work in English at the University of Delaware.
Connect with Susan at firstname.lastname@example.org or @sjaymiller.