Arun Balajiee Lekshmi Narayanan is a Ph.D. student in the Intelligent Systems Program at the University of Pittsburgh. He researches Natural Language Processing and Applied Artificial Intelligence in Education.

Research Interests: Educational technologies; AI in education; cognitive psychology

Fun fact: I love to go on long walks and hikes in dense forest-like parks in the city.

How did you get started in doing work in emerging technologies for teaching and learning?

When I started my Ph.D. in the Intelligent Systems Program at the University of Pittsburgh, my previous advisor, Dr. Erin Walker, oriented me toward this space of research where computer science intersects with learning sciences. Given my background in computer science, it seemed an apt space for me to apply the principles of developing innovative technologies within learning spaces where there is a growing need to innovate. Specifically, my work currently focuses on developing intelligent reading system technologies for undergraduate students majoring in computer science with Dr. Peter Brusilovsky.

What is your most recent insight from your project on emerging technologies for teaching and learning? What are 1-2 challenges you are facing?

My current project involves the development of a reading system that can be used in Introductory programming classes and classes that involve reading research papers for first-year graduate and undergraduate students. From the data collected so far and our analysis of the data that we collected with the reading system, students generally participate in the activities that we provide in the system only up to the requirements of the course that are graded. Although there are several students who go beyond the course requirements to leverage the benefits offered in the system, such as exercises and external web resources, there are several challenges in the space of emerging technologies for teaching and learning. Some of these challenges include 1) adaptability/personalization of these technologies, 2) the considerations of ensuring student/teacher/family privacy, and 3) user control of their data. The idea of user control and adaptive personalization, also known as user modeling, has been explored by researchers in our group, but this would still require possibilities to generalize beyond the limited scope that we consider in our research.

1. Akhuseyinoglu, K., Milicevic, A.K., Brusilovsky, P. (2022). Who are My Peers? Learner-Controlled Social Comparison in a Programming Course. This work discusses the ideas of Open Social Learning Modeling in Technology Enhanced Learning Environment for Programming in Python. The study was conducted with an undergraduate computer science class with 174 students in a large Australian university. Working with practice programming exercises correlated with better course grade outcomes.
2. Thaker, K., Huang, Y., Brusilovsky, P., & Daqing, H. (2018). Dynamic Knowledge Modeling with Heterogeneous Activities for Adaptive Textbooks. This work discusses knowledge/concept modeling based on different quizzes and other reading activities that students do when interacting with an “intelligent” digital textbook in a computer science course. Two models, using Knowledge Tracing, were constructed to assess and predict student knowledge over the duration of the course using the data and logs from their interactions with the digital textbook interface.

What do you gain when you work with others in the CIRCLS community and what are some key lessons you have learned from that collaboration?

I worked with a team during the AIEd Policy sessions and learned many new perspectives to look at while developing innovations that consider privacy and ethics. Specifically, designing elegant practices of literacy in the form of recommendations for teachers/ students/ other stakeholders involved in the use/administration of AI tools in education. The other aspect that I explored with others was understanding the awareness of the existence of these technologies among the various stakeholders who will be using these tools such as educators and practitioners, as well as policymakers as the directives in the manuscript were written keeping this target audience as readers. This was the first of its kind of white papers/articles that brought together/collated different aspects of important topics such as data privacy, stakeholder politics and AI related issues that need to be kept in mind for people who are not tech-savvy. I also learned the importance of informing instructors of the strengths and limitations of education technologies along with plausible policy recommendations that could be made based on this understanding.

Final thoughts:
I’ve attended several sessions hosted by CIRCLS and Digital Promise, particularly the Emerging Scholar CIRCLS sessions. I’ve had the opportunity to network and chat with people who have different levels of expertise and experience, who are practitioners and instructors as well as researchers developing tools to enable innovation in education. I’ve gained several insights which have helped me with my research and increased my awareness of the complexity of incorporating and deploying AI Education technologies in practice.