$$News and Reports$$

Nov. 18, 2015
 

Dr. Kobi Gal of BGU and Prof. David Karger of MIT have collaborated on a research involving Karger’s invention,  NB (nota bene) which is a web-based collaborative annotation tool that facilitates communication among students and their instructors, centered around better understanding of course reading material.  Thousands of students use the platform in dozens of courses all over the world. 

Karger, who is head of the Haystack group at the Computer Science and Artificial Intelligence Laboratory at MIT, says "There's lots of experimental evidence that substantive discussion among students is the very best way for them to learn.  Online tools offer an appealing opportunity increase and enhance the role of discussion in learning in both online and traditional learning, if we can figure out how to use them properly." 

Gal, head of the Human-Computer Decision-Making Lab at BGU says, “There are hundreds of thousands of comments in the forum but no one was looking at how the students were using it. By analyzing posts from previous courses, we can predict where teachers are likely to need to intervene and clarify for the students. We can predict the threads that will generate confusion ahead of time and make this information available to teachers.” 

The project was undertaken as part of the BGU-MIT Seed Fund launched last year, which was organized under MIT’s MISTI program which connects MIT students and faculty with research and innovation around the world. 

Gal and his students were uniquely suited to the collaboration having developed the first online plan recognition algorithm that is empirically shown to infer students’ use of open-ended educational environments. They have also augmented traditional educational software with algorithms for personalizing educational content using machine learning. 

Gal and two students, Orel Elimelech and Avi Segal, spent a few weeks in Cambridge MA this summer working with the Karger lab. They mapped how the students used the forum and used machine learning to predict the expected thread length from an initial post in the forum, which is a proxy for student confusion about the topic at hand. 

“There were 128,000 single comments – that means that it’s likely those students did not get the answer they were looking for,” he surmises. By analyzing the data, the course instructors can give the students better service through the suggestions of when to reply and intervene.  

Gal and Karger are currently writing a joint article based on the research study’s findings. 
Elimelech.jpg
Above: Orel Elimelech in Cambridge

For Elimelech, the experience was valuable both academically and socially. "It was a great experience to work with scientists who lead the computer science and artificial intelligence research worldwide. From day one in Cambridge I felt surrounded by so much talent, and it motivated me to learn and explore as much as possible. Moreover, the student life in Cambridge - hanging out in Harvard Square during the evening and going to a pub - was amazing. I am grateful for the opportunity given to me, and I'm quite sure I'll visit there again in the future," he says.