Hands-On Intro to Machine Learning

Photographs, taken by field scientists, tourists, automated cameras, and incidental photographers, are the most abundant source of data on wildlife today. There are increasingly Machine Learning and computer vision methods that can automatically identify not only species of animals but down to individual animals. That allows us to use photos for wildlife research and conservation. Even social media photos!

Are social media photos a good source of information for science, conservation, and policy?

Can we automate the use of social media images for wildlife?

Let's take a look at a social media site, Flickr(TM), and try to find out what native species of animals are there in Columbus.

• We can use API to download photos corresponding to a query. How would you structure the query? What are the challenges?

• Once we have the photos, we can run a species classifier on the set. How accurate is it? What are the problems?

• What species did you get? How does it compare to the list of species known to be in Columbus? (we'll get the list)

What are your conclusions? Can you think of a way of improving the process? (Technology, bias, incentives, gamification)

Prof. Tanya Berger-Wolf will lead the above discussion and hands-on research activity.

Participants will be briefly introduced to exploreCSR@OSU, which encourages undergraduates, especially students from underrepresented groups, to get involved in CS research. exploreCSR will have several events throughout the academic year.

When: November 17 @ 6 pm

Where: On Zoom (link will be sent to registrants)

Who: Any undergraduate student interested in computing research. Students who are members of underrepresented groups are especially encouraged to participate.

Register here