Ph.D. Related

Posted on Edited on In

3 qualities of successful Ph.D. students: http://matt.might.net/articles/successful-phd-students/

10 reasons Ph.D. students fail: http://matt.might.net/articles/ways-to-fail-a-phd/

The Ph.D. Grind: https://www.goodreads.com/book/show/15731248-the-ph-d-grind

Faculty Job Search: https://pg.ucsd.edu/guo-faculty-job-search.pdf

Advance: https://pg.ucsd.edu/early-stage-PhD-advice.htm

SoK Paper: Oakland

image-20210205094645154

image-20210205101124613

1. Conference Ranking

CCF Ranking: http://harperchen.qiniudn.com/%E4%B8%AD%E5%9B%BD%E8%AE%A1%E7%AE%97%E6%9C%BA%E5%AD%A6%E4%BC%9A%E6%8E%A8%E8%8D%90%E5%9B%BD%E9%99%85%E5%AD%A6%E6%9C%AF%E4%BC%9A%E8%AE%AE%E5%92%8C%E6%9C%9F%E5%88%8A%E7%9B%AE%E5%BD%95.pdf

Tsinghua Ranking: http://numbda.cs.tsinghua.edu.cn/~yuwj/TH-CPL.pdf

Third-pass approach

  1. First pass

    Steps:

    1. Carefully read the title, abstract, and introduction
    2. Read the section and sub-section headings, but ignore everything else
    3. Read the conclusions
    4. Glance over the references, mentally ticking off those you've read

    Aims:

    1. Category: What type of paper is this? A measurement paper? An analysis of an existing system? A description of a research prototype?
    2. Context: Which other papers is it related to? Which theoretical bases were used to analyze the problem?
    3. Correctness: Do the assumptions appear to be valid?
    4. Contributions: What are the paper’s main contributions?
    5. Clarity: Is the paper well written?

    Results:

    1. Choose to read further or not

  2. Second pass

    Steps:

    1. Look carefully at the figures, diagrams and other illustrations
    2. Mark relevant unread references for further reading

    Decide one of the following further action:

    1. set the paper asidee
    2. return to the paper after reading background material or
    3. persevere and go on to the third pass.

  3. Third pass

    Steps:

    1. Make the same assumptions as the authors, re-create the work.
    2. Compare this re-creation with the actual paper
    3. Identify a paper’s innovations
    4. Identify hidden failings
    5. Identify and challenge every assumption in every statement
    6. Think about how to present a particular idea
    7. Jot down ideas for future work

    Aims:

    1. Reconstruct entire structure of paper from memory
    2. Identify strong and weak points
    3. Pinpoint implicit assumptions
    4. Pinpoint missing citations to relevant work
    5. Pinpoint potential issues with experimental or analytical techniques
  1. Use well-chosen keywords to find three to five recent papers in the area.
  2. Do one pass on each paper to get a sense of the work, then read their related work sections
  3. If you can find a recent survey, read it.
  4. Otherwiese, find shared citations and repeated author names in the bibliography
  5. Download the key papers and set them aside
  6. Go to the websites of the key researchers and see where they’ve published recently
  7. Identify the top conferences in that field.
  8. Go to the website for these top conferences and look through their recent proceedings
  9. These papers, along with the ones you set aside earlier, constitute the first version of your survey.
  10. Make two passes through these papers. If they all cite a key paper that you did not find earlier, obtain and read it, iterating as necessary.

  1. Learn to read papers and develop your taste

    1. Relevant seminar class
    2. No need to understand all technical details of a paper
    3. Appreciate and criticize research ideas How to read a paper
      • Why is a paper good or bad
      • What makes an interesting paper

  2. Recognize patterns of developing research ideas

    1. Pattern #1: fill in the blank Systemization of Knowledge (SoK)

      Jot down notes about the differences among these papers in terms of assumptions, guarantees/properties offered by a system, methodologies, techniques, datasets, etc. and then draw a table. Look for empty spots and those are potential new research ideas you can work on. You can make a much more fine-grained table which will give you more opportunities to identify the blanks.

    2. Pattern #2: expansion

5.1 Steps of doing a review
  1. Print them out, write paper numbers on the first paper
  2. Read each paper in turn carefully
  3. Scrib notes in the margins with whatever comments you think of at the time
5.2 Structure of a review
  1. Summarize the paper
    1. Give a neutral description of what you think the paper is about
    2. Where the authors are coming from
    3. Why they view the problem as important
    4. What they’ve done
  2. State what you think the contributions are
    1. Is the contribution useful
    2. Is the contribution flawed
    3. Point out if the authors fail to state the contribution
  3. Give specific comments on the paper
    1. Small points link typo => nits
    2. Novelty
      1. what's new about the work
      2. Is there some related work that the authors have missed?
      3. Does the related work invalidate the contribution, or (more likely) simply change it’s context or emphasis?
    3. Written
      1. How well written is the paper?
      2. Could it be made clearer? Suggestions here range from running a spell checker or improving the language, to rearranging entire sections of the paper to make it flow better.
    4. Shortcomings?
      1. Are there gaps or unaddressed issues?
      2. Are there any apparent technical flaws?
    5. Advantages?
      1. Was there anything you thought was really cool about the paper?
      2. Is the paper likely to prompt interesting discussion at the conference or workshop?
      3. Is the paper appropriate for the venue?
  4. Conclude the paper, give a brief recommendation for the paper and your reasons for it.
5.3 Tone

Several examples

This system doesn’t deal with unexpected vegetables ==>

The paper would be much stronger if it discussed how the system deals with unexpected vegetables.

This paper doesn’t cite Multics, which did everything you do and more ==>

This paper reminded me of Multics, which seems quite similar. I would find the paper more persuasive if it stated what the authors do over and above Multics.

The algorithm given in the paper breaks in the presence of Byzantine faults ==>

The description in the paper left me worried that algorithm breaks in the presence of Byzantine faults.

6. Misc

一点实用技巧：

1 不如开始做。如果只给你25分钟，你怎样让工作成果表现得尽量好看？就要这个“虽然不完美，但也比没有强”的结果。用25分钟去做。如果还有困难，25改成5分钟。总可以做一些事情的，等25分钟做完以后，再问自己这个问题“再给我25分钟，怎样让工作成果尽量好看？”。

2 立刻沟通。如果你的 deadline 是跟别人承诺的，立刻告诉他你现在的进展（没有进展也要告诉他），然后，就在面对面，或者电话连线这种临场环境下，估计一个最快给他看进展的时间——不是最终结果，而是“总比没有强”的进展，如果能半小时给，就不要拖到一小时。

（是先做？还是先沟通？取决于你对别人有没有承诺。如果纯粹是自己给自己定的 ddl，就先做。其他情况下一定要先沟通，用电话或面谈，双向沟通，不要用留言的方法。）

背后的原理：

3 你的焦虑、ddl、压力什么的，其实都是浮云，都是狗屎——用不了多久，也许下周，也许明年，它们就过去了，可能回想起来只是个笑话而已。此情可待成追忆，只是当时已惘然。不要盯着这些浮云，你是什么样的人，是你最终做了什么事决定的。所以还是回来盯着那个事情。

4 虽然是浮云，但是它们会让你很不爽。因为你“在做”和心里认为“该做”的事情有很大距离。所以我说“不如开始做”，只要这个距离缩小了，不爽也就解决了。当进度出现问题的时候，最该做的就是第一时间告知队友和老板/老师，比开始做更重要。

5 不要以为告诉“没有进展”是很丢脸的事情。因为对方也是人，他也会拖延，你没有进展他也会容忍的。你最激怒他的是拖延了而不告诉他，因为这是欺骗。就好像开车误撞人以后，怕对方索赔就把人杀了，然后毁尸灭迹的过程中被人发现，然后杀人灭口，然后被警察追杀又杀了警察…… 拍成电影很刺激吧，这就是你说的恶性循环。在一开始就戳破这个循环！

6 培养让自己更舒服的工作习惯，比如一份远程工作，在早晨，留言告诉对方我要在今天达到什么进展，在晚上，告诉对方完成情况，如果没完成，什么原因也写清楚。早晚花10分钟，一天都没有压力。团队合作中这样的习惯尤其重要，培养习惯在刚开始不容易，要先明白一件事：这些习惯要跟着你五十年，如果是让你舒服的习惯，就舒服五十年；如果是恶性循环，就恶性五十年。宁可要舒服对不对？

7 关于好习惯，其实有很多规则和技巧。但你可以从我的1、2开始。提高效率是五十年的事，不要临阵磨枪去学时间管理，一边还扔着拖延的工作。沟通，干活。

img