acm-header
Sign In

Communications of the ACM

ACM TechNews

An AI for Deciphering What Animals Do All Day


View as: Print Mobile App Share:
The hydra's neurons are labeled with a green fluorescence indicator.

Columbia University researchers have demonstrated how an algorithm for filtering spam can learn to pick out, from hours of video footage, the full behavioral repertoire of tiny, pond-dwelling Hydra.

Credit: Yuste Lab/Columbia University

Columbia University researchers have demonstrated how an algorithm for filtering spam can parse hours of video footage to reveal the behaviors of Hydra, a close relative of coral, jellies, and sea anemones.

Although it lacks a backbone or brain, Hydra behaves in predictable ways that computers can identify.

The research aims to illuminate Hydra's nervous system functions by comparing behaviors to the firing of neurons.

The team applied the popular "bag of words" classification algorithm to hours of footage tracking Hydra's actions. The algorithm identified 10 previously described behaviors, and determined the impact of various environmental conditions on six of those behaviors.

The researchers plan to decipher Hydra's neural code with a model that shows how its networks of neurons create behavior.

Beyond facilitating behavioral studies in more complex animals, the work has potential for maintaining stability and control in machines that navigate in variable conditions.

From Columbia News
View Full Article

 

Abstracts Copyright © 2018 Information Inc., Bethesda, Maryland, USA


 

No entries found

Sign In for Full Access
» Forgot Password? » Create an ACM Web Account