Stephanie Milani will complete her Computer Science (B.S.) and Psychology (B.A.) degrees in 2019. She is a Grand Challenge Scholar, a URA Scholar, and a member of the Honors College.
Title of your research project: Creating a Scalable Framework for Model-Free Reinforcement Learning in Norm-Rich Environments
Describe your project:I developed a scalable framework for robots to incorporate human values and preferences, or norms, with traditional methods for learning how to perform tasks so that these robots can learn how to act in environments where there may be social ramifications to actions.
Who is your mentor for your project? Why did you choose them? My research mentor is Dr. Katia Sycara, Robotics Institute at Carnegie Mellon University.A couple of my friends from the Robotics Institute Summer Scholars program worked with her last summer and only had great things to say about working with her. When she and I met, I liked the research questions that she proposed. Based on that conversation and the research publications I had read from her lab, I decided that I wanted to work with her for the summer and the fall.
How did you become interested in this project?Dr. Sycara's lab is researching the general area of normative reasoning for social robots. While reading papers in this area, I realized that few people have looked into how to incorporate normative reasoning in a reinforcement learning setting. I've been interested in reinforcement learning for a few years now - both from its psychological underpinnings and in the area of machine learning - and I wanted to take the opportunity to explore it in a little-researched area with real-world impact.
What has been the hardest part about your research/what was the most unexpected thing about being a researcher? The most difficult part about my research is finding a research problem that is both specific enough to make a tangible contribution and plan, but broad enough that it has greater impact outside of its original context.
What has been the most rewarding part?Probably presenting my research and being met with interest!
How will you disseminate your research?I will publish this research in the RISS Working Papers Journal. I plan to incorporate part of this work in my URCAD presentation, as well as other publications.
What is your advice to other students about getting involved in research? If you find someone's research findings interesting, then contact them about it! They may have opportunities for you to work with them.
What are your career goals?I plan to obtain a Ph.D. in Computer Science, specifically focusing on the field of reinforcement learning. After that, I plan to continue with research in machine learning in either an academic or industrial setting.
To learn more about becoming a URA Scholar:
or contact Dr. Householder at : firstname.lastname@example.org