Control, Optimization, and Online Learning for Autonomy Lab

Q&A FOR PROPSPECTIVE STUDENTS

The goal of COOL Autonomy Lab is to create a supportive and fun research environment for students to learn and develop necessary skills for their future careers. We are always interested in working with highly motivated students having a strong mathematical background. If you are passionate about designing the next generation of control, optimization, and learning algorithms for autonomy, feel free to reach out to me. Below are some guidlines and information on what I expect from you and vice versa.

PhD Students

1. What is my goal for PhD students?

I always believe in a strong foundation. That said, students will have opportunities to learn fundamentals that will enable them to tackle their research problems. When they graduate from our lab, they should have not only a good foundation but also deep understanding and some results of one research area that they choose during their PhD. These intial results will hopefully enable them to be successful wherever they will end up with.

2. What type of research do I focus on: Theoretical vs applied?

Our research interest lies on the intersection of theory and applications, i.e., develop new efficient methods with theoretical guarantees, simulate the proposed algorithms, and then test them on real hardware, e.g., robots. Ideally, we want to have a balance between theory and real implementation.

3. What do I expect from incoming PhD students?

Students joining my group should feel comfortable to take theory classes in control, optimization, and probability from AE, ECE, and Math departments. While taking courses (during your first year), students should explore different research areas and should know what problems you want to solve. Once you have sufficienly finised courses for your degree (probably in your second year), you should fully focus on addressing challenging research questions in your desired areas. That said, I expect that students will be selfmotivated and indepedent in their research. This is your degree. I already finished mine!

4. How often will I interact with PhD students?

I am very hands-on person or at least I try to. I’m always interested in learning new things, and one way is through my students. I always want to meet with my students at least one per week to get an update on what they have been working on and how can I help. I believe that to be succesful, students and their advisor should communicate efficiently with each other. Here are a few examples that I don’t want to have.

1) Ghosting is extremely bad: It can mean you don’t work and don’t have anything to discuss or we don’t like each other and can’t talk. So, if we schedule a meeting, I will definitely not ghost you and expect you do the same.

2) You disappear with notice and only show up in a few weeks later even with a nice result, I don’t like it. If you can work on your own then why you need me.

5. How many papers do I expect from PhD students per year?

I do not have a golden number. I think students should publish one or two conference papers per year (from their second year) that will enable them to attend conferences to network with their peers. On the other hand, I prefer journals because we have to go deeply to our research to have sufficiently good results to get accepted.

6. What venues do I submit my papers?

I always try to submit my work to well-known and well-recognized conferences and journals. There are two reasons: 1) I will likely get good reviewers to judge my works (although this is likely not true for ML conferences); and 2) I can often network and learn from top-notch reseachers in my area. Few examples of these venues include CDC, ACC, COLT, IEEE TAC, SIAM journals, etc. I always encourage my students to do the same. However, students will lead the effort of these submissions.

7. How do I decice when a student can graduate?

I do not decide when a student can graduate. There is a committee from the department for that. To some extend, my role, as your advisor, is to tell you my opinion on your progress. At the end, you will be the one to decide whether you are ready to gradute, I think. In general, you should make sure that you will not finish your PhD in 7 or more years, which is not common in engineering.

8. How I finacially support PhD students?

Students joining my lab will be supported under Graduate Research Assistant (GRA), which comes from my research projects. In general, I will guarantee that the students will be on GRA for about two years. Funding is always unpredictable but I always try my best to support my students. I also encourage my PhD students to have at least one semester of doing teaching assistant (TA). This can help students to improve their communication and teaching experience.

9. How to apply to our lab?

The best way is to apply to the PhD program in our department and mention my name as your potential advisor. I encourage you to check out our recent publications and make sure that our research and culture align well with you, i.e., how we can help you to develop your career and what will you bring to our lab? If you have any questions/comments for our research, feel free to send me an email.

MS Students

To be updated.

Undergraduates

To be updated.