Tips on Ph.D. study

Tips on Ph.D. study

1. Ph.D. Dissertation vs. M.S. Thesis:
    The differences between a Ph.D. dissertation and an M.S. thesis are in the degree of (1) depth, (2) formality, and (3) roadmap. For depth, your problem should be more challenging or forward-looking than the one in an MS thesis. For formality, your solution might be an optimal one instead of a heuristic one. That is, you need to give a “formal” treatment to a well-defined problem. For roadmap, often each chapter is a research paper, and all chapters together form a roadmap. A roadmap is a series of design, implementation, testing, and analysis, or a series of related topics.

2. From Ph.A. to Ph.D.:
    There exists 3 checkpoints within your Ph.D. study. You start as a “Ph.A." when you are admitted to the Ph.D. program. After you pass your qualifying exam, you become a “Ph.B." This should be done within one year, at most two years. After you present your dissertation proposal to the department, you then become a “Ph.C.” This should be done after you finish and submit your “first” paper. After you full-fill all requirements with a dissertation along with some publications, you turn yourself to a “Ph.D.” A quick hand could finish the entire process in 3 years, but a slow hand could take 7 years or more. The average is something around 5. You should set a goal to finish in 3 years, though you might end up 4 years if having some delay. The key is to work with a tempo and a sense of urgency.

3. Acceleration:
    An M.S. student has clear peers to synchronize with, and thus has enough pressure to graduate in two years. However, a Ph.D. student does not have such a peer pressure, and often has floating schedules. How long to graduate thus depends on your self management. Nevertheless, there could be two ways to clock yourself faster: (1) conference-driven, (2) peering. You could target at the submission due date of a conference to push yourself. In principle, a Ph.D. student should have at least two experiences of paper presentation in a conference. But if the same paper is to be submitted to a journal later, an extension of over 30% is required. Alternatively, you can look for the due date of a journal’s special issue that fits your research. Another way is through peering with another Ph.D. student, M.S. student, or post-doc to co-work on the same paper. If peer A hosts paper X and helps on paper Y, while peer B hosts paper Y and helps on paper X. As a pair of peers, they could trigger each other to push things forward.

4. Research Cycle:
    A paper starts from an idea which is then consolidated to a proposal. A proposal should contain (1) motivation, (2) background, (3) issues, (4) problem statement, (5) survey, (6) solutions (or algorithms), (7) evaluation plan, (8) expected contributions, (9) schedule, (10) references, where problem statement is the most important part with given inputs, desired outputs, and constraints, all defined in parameters. Note that a problem well defined is a problem half solved. After the proposal, you design, implement, analyze, and evaluate the solutions, and then write up the paper. Before writing up the paper, it is better to have an outline to lay out the logic flow. An outline should have (1) section titles, (2) subsection titles, (3) paragraphs in keywords to list the logic flow and arguments, (4) numbered titles of inserted figures and tables, (5) numbered references in [ ]. (6) list of references. With an outline at hands, you can concentrate on sentence composition without stopping to think about logic flow. Usually you can finish writing up a paper draft in one month if you have a detailed outline. From idea to submission of a paper, one should manage to finish within one year to avoid losing interest and passion on that topic. From submission to acceptance to a journal or magazine, it ranges from 6 months to one year or more, depending on efficiency of the journal/magazine your editor, and how many rounds of revision. From acceptance to publication, it could be several months to half a year.

5. Publication – 1st-tier vs. 2nd-tier:
    Another way to graduate soon is to submit papers only to 2nd-tier or even 3rd-tier conferences and journals which have lower quality standards. But the likely visibility and impact of your work would be reduced. So is the creditability of your track record. You should target at the 1st-tier. If it gets rejected, you consider another 1st-tier or switch to 2nd-tier. If your work is on a very new topic but not deep enough, you could submit to an IEEE magazine. If your work is deep and solid research, try an IEEE transactions or a top confernce. For something in between, you try the other journals or conferences.

6. Technical Contributions vs. Paper Engineering:
    A paper with great technical contributions but poorly written would go no where. But a paper with medium technical contributions but very well written could be accepted at a 1st-tier target. While you pursue great technical contributions as the main goal, enhancing your skills of “paper engineering” is a must. Paper engineering in academia is similar to “product engineering” in industry. If you cannot do well for one of them, you probably cannot do well for the other. Rigidness is the key here. The art of paper engineering includes (1) tight logic flow, (2) flawless grammar, elegant sentences, and precise words, (3) well formatted style. Most non-native speakers of English think that (2) is the main barrier. But actually (1) accounts for 60% of the writing problems. For logic flow, always remember WHY is more important than HOW, and they should interleave with each other. Give logical arguments instead of plain descriptions. On the other hand, though a well-written paper with medium technical contributions could get accepted in a 1st-tier target, its impact in terms of citations would not be high. Often the impact of a paper is half decided when the problem statement is written down. The other half depends on your solution. Remember that a new problem with an old solution usually has higher impact than an old problem with a new solution. Thinking about the problem and its solution is the most interesting part of your research activities. If you did not have the experiences of pondering on a problem or a solution for weeks or even months and suddenly you came up with a brilliant idea, you have not done real research yet. That is the magic moment of your research that you will never forget.

7. Research Only vs. D&R:
    As a Ph.D. student, you might think that your job now and in the future will be pure research and no development. Then you probably would pick your problem from the literature survey and lose your development skills during your Ph.D. study. The former would likely lead you to work on an old problem instead of a new one, and hence with lower impact. The latter would cornor yourself to an academia job. You dare not go to the industry, period. Instead, you should equip yourself with both research skills and development skills. You could dive into development to find real problems for your research. In that sense, research is the non-trivial part found during the development process. Your chance to find a new problem is also higher.

8. Academia Job vs. Industry Job:
    Once you have both skills of D&R, whether to get an academic job or an industry job depends on (1) personal interest, (2) job opportunity. It wouldn't depend on your capability set because you got both D and R skills. Knowing the differences between R and D and how to merge them would put you in a better position, no matter in academia or industry.