GRADUATE STUDENTS

EXPECTATIONS, RIGHTS, ADVICE

 

Here are some advice and information about rights and expectations that  especially starting grad. students may find useful.

More senior students may benefit from presentation and paper writing tips I give over and over to my students:

 

Choosing your advisor, topic, and starting your research:

1.      Choose the right topic: I suggest that you work in an area that you find fun and easy for you. It is much easier to be successful and happy in the future if you like your area. The courses you take is a good indication I would say. Also allow yourself about a semester to decide.

2.      You may change your advisor/topic: I hear many students who think that it is wrong or inappropriate to change their advisors or their topics. This is clearly your right (things may be more complicated in terms of your support when you are working on a project, but even then, it should be possible). However, is important is that you talk about your concern or decision to change first with your advisor. What is not proper is when your advisor hears about the change after the fact.

3.      Publish or perish: If you choose an academic career, publication is the number one criteria for success. You should start building that skill and name recognition starting with 2nd or 3rd year of PhD studies.

4.      Presentations: See presentations as opportunities to show your work to people in your field, which means you should try to do your absolute best. Even if it may seem like extra work, doing practice talks will get you valuable advice and practice.

 

Graduate Student Expectations:

I have put here my expectations from graduate students and TAs working with me, but they are pretty much standard expectations.

  1. Be available at school most of the work time: One of the main issues is that the school is far from the city and the transportation is expensive. Nonetheless, in return to your scholarship, whatever it may be (FS, TW, Project or BIDEB), you are expected to be at work (SU) at least 3-4 days a week, during school period. Communication over email is not only inefficient, but being among colleagues will improve your research (you can get more advice from your advisor and help/feedback from fellow students). You will also contribute to research life by giving feedback about others’ work. Being at school most of the time  is also very important for Taship duties: certain TA duties that cannot be planned ahead and we rely on students who are around to help with such situations (e.g. problem in homework description).
  2. Be timely: If you are expected to do something at a certain time, please do it on time or email your professor about when the work will be done –well before the deadline.
  3. Respond to email promptly: It is unacceptable that you email a professor, and then s/he responds not to hear from you again until the next day, during work days. You should be reachable typically within half a day.
  4. Improve your English: Much of the academic work is writing. So the sooner you improve your writing skills, the better. For that, try to attend a writing workshop, read and learn from well-written papers, read a book about common mistakes (I have a good one)...
  5. Keep a research log/book: Many times the student come to my office and at the end of the talk they ask for my notes to take. Keep a book to take notes and progress and log experiments, it will be very useful.

 

TA Expectations

1.      Try your best: TAship should not be seen as a burden, since its part of your job. You should see any job that you do from now on as an opportunity to develop and show yourself as a good worker/colleague. I for one, and many others, would rather work with someone who is studious, timely, and responsible than someone who may be somewhat better –academically.

2.      Be timely: If you are expected to do something at a certain time, please do it on time or email your professor about when the work will be done –well before the deadline.

 

Writing good papers and Presentations:

Many students would really benefit from a technical writing class; but the following are just short tips in case you did not/could not attend one yet.

 

Writing papers:

1.      General to specific/First to Last: Clear your mind and imagine explaining your work to someone unfamiliar with it; starting from the general and going to the specifics and first steps of your work to later etc, in order.

2.      One idea per paragraph: You should use have one idea/topic per paragraph and explain ideas in enough length/detail.

3.      Smart sentencing: If you are having too much difficulty explaining something and your sentence gets long and awkward, try doing so in a few lines. On the other hand, your writing may be unnecessarily choppy too: if two sentences are connected, make that connection concrete (use however, on the contrary,...) or use a semi colon.

 

Preparing presentations:

What is important in a presentation is that it arouses interest in people, so that hopefully they will remember you and your work after the presentation. Hence, we can say that a presentation that does not elicit any question is a bad one. Also remember that this is your chance to impress people who may be future supervisors, colleagues, boss etc. So here is a quickly written list about how to make it a good talk:

 

1.      Timeliness: Be timely! One indication of a person who does not plan well or adapt himself is that s/he is not timely (to finish). Reserve 2-5 min in a 15 min presentation for the intro according to the composition of the audience. Make sure people understand the difficulty of the problem at hand. Another 3-5 for results and the rest for the method. If time seems to be short, skip some parts with a brief summary. Dont talk too fast to cover more things, rather a slower talk is better - so skip some details. The only exception is intro or the important part of your method so that you draw attention.

2.      Level of detail: Most/all people should understand the topic and why the problem is difficult; many people should understand your approach and some people should understand and be able to criticize your method. Do not give too much detail that people loses interest, they can always go and read paper. It is often better to remain on the simpler side (in the sense of more global view) than to loose people with unnecessary or unsuitable details.

3.      Slides: I would prepare 20-30 slides for 15 min. I like them dense (figure, at least 5-6 items) rather than 1-line slides and too much transitions. This lets people who like to see the whole thing more freedom and looks more sophisticated. I also prepare some slides on parts that I dont plan to present but I may use if a question is asked.

4.      Segments of talk: I make sure I know how to direct the talk in some important slides. So you may study what you want to say more ahead of time. I also try to give some summary before starting a new segment (e.g. “now that we have extracted the features, the task is to train a classifier”...)

5.      Questions: Neither dodge questions (e.g. I dont know.") nor be snob about answering them. How you answer questions may be as important as the impression you gave during your speech. Finally, err on the humble side and give credit for the question (e.g. that’s a good question) whenever appropriate. Use criticism/questions to improve your work.

6.      Exciting, interesting talks: Use stops in your talk to highlight important parts, so that people understand and appreciate the issue. Maybe ask a question, let the audience think very briefly, and then you answer