Research Guide

Here is a very short, concise beginner guide for AIT Master students (applicable to Ph.D. students too).

First step to choose topics is to visit Projects. I TRY NOT to work on any other topics outside of this list.

Graduation criteria

To graduate from AIT Brain lab, Master students should either (1) develop a workable, deployed product (justification of 1-year effort) with formal user evaluation, or (2) publish in top-tier journals / conferences. Ph.D. students should publish 3 top-tier journals / conferences.

Meeting protocols

For Master's, we will have weekly meetings. For PhD's, it is monthly meeting. In the meeting, you briefly revisits your research questions and objectives, talk about progress, bottlenecks or obstacles that require attention, and next steps.

I personally do not prefer scheduling separate meetings as my schedule is very unpredictable and thus could hurt your planning. Try to talk as much in the regular meetings. But if really needed, check my free slots in chaklam@ait.asia and book accordingly.

Advices

Research can be very difficult. Here are my distilled short advices. You should

  1. understand that research is not linear
  2. prioritize progress, over perfection
  3. read top-tier papers
  4. validity, then novelty, then clarity
  5. make your research questions and objectives clear
Useful links

Character

It's well grounded that research is not only about intelligence but also of your attitude and mindset. Here is what I think my successful students do, compared to less successful students.

  • Strong passion: this is perhaps a very important factor separating the successful and the less ones. Passion is not about career, money or for fun. Passion is something you truly believe in. If you really love the topic you are doing, the rest of the below factors come more easily.
  • Self-discipline: keeps you focus on tasks on hand and constantly thinking about your work. No single successful students do not have self-discipline because research is difficult.
  • Plan well: they make a good plan first instead of just blindly doing things. They execute the research in a well-planned systematic logical manner.
  • Critical thinking: they question everything and challenge the assumptions. This will yield good questions to solve.
  • Creative: they are good in linking and synthesizing and creatively come up with new, original idea.
  • Positive attitude: they seldom blame anyone if they fail. They keep going, know how to positively self-talk, and embrace and accepts struggles as learning process. Under huge stress, their positive attitude carry them forward.

1. Timeline

Here is a simple figure explaining your one-year thesis journey. You have one year to complete your thesis; most students extend to one or two more semesters:

timeline

2. Writing

Writing is a good way to organize your ideas. Try use this template to communicate your ideas effectively.

  • BACKGROUND: The ability to differentiate person in dialogue summarization is important. Past work mostly focuses on ________ (talk about what people mostly do…)
  • PROBLEM: However, the key limitation is ___________________________(major gap in past work)
  • SOLUTION: This paper proposes _____________________ (solution that directly fix the gap)
  • RESULTS: We found that ___________________ (results that confirm we fix the gap…)
  • CONTRIBUTION: The results implied that _________________ (generalizability of your work…)

3. Do your research!

Doing your research is a daunting task. I have provided a very short snippets of success you may want to try for both fundamental research and applied research.

how to do research

how to do research

4. Deliverables

Your main deliverables for your thesis are

  1. the public GitHub link of your code,
  2. final thesis report
  3. video demo (if you are applied research),
  4. final publication (if you are fundamental research)