PhD Students

First-year Rotations

Rotations are really short, so the way we do rotations in our lab are either: 1) Brainstorming on a new project for a couple of weeks, and working on that; 2) Joining an existing project and helping out one of the grad students in the group. #2 is usually more successful because the project is more well-defined and can be tackled in the span of a quarter. So I recommend the rotation students to talk to all members of the group and discuss this possibility with them. #1 is also possible, but since rotations are only 10 weeks, we should be very realistic and define a well-formed problem that could be done in the span of 10 weeks. You should try to finish your project during the rotation period. If the project is not finished and there is a non-trivial amount of work that needs to be done, the project could be transferred to other students in the group and depending on your contributions you might or might not be involved in the final results and the paper. If most of the work is done during the rotation period, and you are willing to still actively lead (spend a non-trivial amount of time on the work when needed and be the first author of the work), you can continue with the project until we have shown results or submitted to a conference/journal. This is often the case for Fall rotations, when students submit their work to conferences such as RSS or ICML which end up being in late January. Please discuss this with faculty who you’d be rotating with in Winter. Faculty (including myself) are often very understanding of this. Winter/Spring rotation projects often make it to deadlines such as NeurIPS or CoRL.

In general, during your Ph.D. you should be working on many exploratory ideas that don’t necessarily lead to a publication, but it is not suggested for a rotation project (you should discuss this with Dorsa if you think this is a better option). The reason is that rotations are very short for just exploring ideas, and it works a bit better to have a very well-defined project idea and goals. This allows us to figure out how well we can work together under a normal project, and have an easier time finding out if ILIAD is a good fit for you. More exploratory projects are doable if you are planning on working on it for more than one quarter (for example working with ILIAD for two quarters or doing a co-rotation with our lab and another lab for 2 quarters but on the same project). Rotations are also quite useful to see if there is a character match between you and the group. We are looking for students who actively join our meetings, contribute to the group, help out other members of the group beyond their own personal projects, come into the office, and interact well with the rest of the group. In addition, some of the other factors that are important for joining ILIAD are being genuinely interested and curious in research in the space of interaction (human-robot interaction, human-AI interaction, multi-agent interaction, interactive learning, interactive teaching, learning from various sources of human data).

Co-advising and Collaboration

Dorsa is open to co-advising students only if there is a natural place of collaboration with the potential co-advisor, and there is already a common problem of interest that all of us have discussed and agreed on.

Collaboration with other faculty usually happens naturally, either through taking a class, or a shared grant proposal or project. Please discuss with Dorsa beforehand if you are thinking about being co-advised or discussing collaborations with other faculty.

Internship

You should always check with Dorsa if you plan to do an internship over the summer. You are expected to be around during summers and work on research. However, you can do up to 2 internships during your Ph.D. and only if the internship is aligned well with your research directions, can be focused on research, and there is an opportunity of collaboration with our lab. Please discuss your plans for doing internships with Dorsa before committing or even applying to anything. It is recommended NOT to do internships during the first year of your Ph.D. after the rotation program. That is a good time to focus on building your research direction.

CAship

Every Ph.D. student needs to CA at least two quarters during their Ph.D. This might not be required by the department (e.g. the EE department), but it is an amazing learning experience. You should discuss your CAing plans with me before committing to any positions. Potential courses to CA are: CS333, CS221, CS237A and CS237B. I will probably ask for your help for CAing some of these courses at some point. I do not recommend CAing courses in your first year.

Quals

Depending on if you are an EE or CS student. You’d need to take your quals in your 2nd or 3rd year. You should discuss your plans for when to take the quals with me at least 6 months in advance, and we would discuss who to include on your committee. You’d need to schedule a meeting with each of your committee members to discuss your research directions before scheduling your qual exam.

Vacation

All of us need vacations and time offs. Please let Dorsa know what your travel plans are at least a month beforehand. Our work is usually flexible which is nice, but be reasonable with your vacation periods and timings. Please avoid vacations during major deadlines that you plan to submit to or during major conference times.

Let Dorsa know beforehand if you need to travel for visa renewals, etc.