Can Erdogan

Major: Computer Science        Year: Senior                 Hometown: Istanbul, Turkey

Projecttitle: "Action Selection via Learning Patterns in a Multi-agent System"

When did you become involved in undergraduate research at CMU?  I started working with my faculty advisor Professor Manuela Veloso in Spring 2010 and we focused on another problem that I presented on Meeting of the Minds 2010 with the title "Learning from Demonstration and Chip Kick Evaluation in Robot Soccer".

How did you find your mentor?  I took a course called "Introduction to Artificial Intelligence" in Fall 2009 from Professor Veloso and wanted to learn more about AI and robotics by doing research with her.

How has your idea/project evolved through the academic years?  Looking for patterns under complicated systems is a very complex AI question and I have faced it over and over in different courses such as AI, Computer Vision and other robotics courses. For this project, we focus on a smaller scope of the problem limited within the domain of robot soccer. The idea for the project evolved during my Spring 2010 research with Professor Veloso as it was apparent that my research then was leading to this problem.

What successes or difficulties have you encountered in this project or others?  There are different types of difficulties that an undergraduate student can face in doing research. On my behalf, some of those difficulties were simply related to starting doing research. For instance, for the first time, I had to write code for a project in such complexity and size. Also, it was my first time doing research, looking for related work and continuously searching for different approaches to solve a problem. On the other hand, at the end of the day, it was also very enjoyable to find a genuine solution to a research problem.

If you could summarize your experience in one word, what would it be?  Thrilling