*********************************
There is now a CONTENT FREEZE for Mercury while we switch to a new platform. It began on Friday, March 10 at 6pm and will end on Wednesday, March 15 at noon. No new content can be created during this time, but all material in the system as of the beginning of the freeze will be migrated to the new platform, including users and groups. Functionally the new site is identical to the old one. webteam@gatech.edu
*********************************
shaun.ashley@physics.gatech.edu
Summary Sentence: School of Physics, Nonlinear Science & Mathematics Seminar, Dr. Souvik Mandal - Harvard University
Full Summary: No summary paragraph submitted.
Souvik MandalQuantification of behavior is one of the primary requirements to study animal behavior scientifically. Traditionally, behavior has been quantified by manually observing the focal animal(s) across a spatio-temporal scale and recording the occurrences of behavioral events. These events are generally deduced from the movement of different body parts of the animal, typical body postures as well as its overall movement. While collecting data by manual observation has several advantages, it is prone to disadvantages like human bias and being imprecise. Though modern videography has improved the observation and recording of behavior, extracting behavioral data from these video data remained challenging until now.
In this talk, I will discuss how the recent progress in machine learning tools has enabled me, a biologist interested in social insects, to extract behavioral data from videos. I will talk in detail about such a tool, called DeepLabCut, which tracks the movement of individual parts of an animal with minimal human input. I will end the talk with an example of the application of this tool in my current projects, which is understanding the evolution of cooperation and foraging strategies in the Carpenter ants.
