Fine Grain Data is a series of interviews with senior ACTNext research scientists conducted by our communications intern, Megan Ciszek.

The aim of the series is to provide an opportunity to get to know some of the researchers working at the cutting edge of developing the next generation of tools for learning and assessment.

We invite you to check back frequently over the coming weeks, and many thanks to Michael for helping us kick off the Fine Grain Data interviews.

(M.C.): Were there any early influences that sent you in a particular research direction? Mentors?

(M.Y.) My undergraduate and graduate adviser in Russia had a big influence on my academic choices. My graduate adviser in the US, Dr. Peter Brusilovsky, formed my taste for research and innovative development.

At Carnegie Mellon you crafted statistical models using Big Data, specifically regarding the acquisition of knowledge. In what ways do you see this work as important or impacting the academic community?

I was interested in building Bayesian Knowledge Tracing (BKT) models from Big Data. I was surprised that one would need a cluster computer and an expensive Matlab subscription to do that. My post-doctoral co-adviser, Dr. Geoffrey Gordon from the Machine Learning department at Carnegie Mellon, suggested I write my own code for fitting BKT models from Big Data. I got extended funding for continuing this project [at Carnegie Mellon University] and, after some time, people started asking me if they could use the utility I built. Since it is open source, it is free for anyone. To this day I see references to that package turning up in literature.

You are currently part of a team within ACTNext working on the Companion app. How will your work on Companion facilitate students in learning and retaining relevant skills?

We are at the beginning of the path to build an integrated learning experience for the students. Right now, we are working with a diagnostic model and collecting student data to make the next sprint towards the big goal – helping students learn anytime, anywhere from their mobile phone.

What aspects do you most enjoy about your job?

Building new, powerful tools for students to use. Working with the best and the most motivated people in the industry.

Aside from the Companion app, are you currently working on any particularly challenging or exciting projects?

I am also participating in David Kuntz’s ALTT (Adaptive Learning Theme Team) initiative. This line of work targets an over-arching architecture for adaptive learning. I’m very excited about this work. Also, I’m working with members of the Holistic Framework team towards a data-driven validation of the item tagging.

Are there any particular aspects of your work, past or present, which you believe will have the most impact on learners/students?

Building adaptive systems for everyday use. It is seemingly not that hard once you know how to do that. But to do it right and for a large audience is quite tricky. I am sure ACTNext has the right talent to do that.

Do you view your work as primarily affecting students, or people who want to learn something later in life?

A student is a state of mind, not an age group. Everyone who is learning is a student, me included.

Can you recommend your favorite Russian bands?

Since [the] early 2000’s I’ve liked Zemphira and the band called Kino.