Tips for PhD

2020-03-01 219浏览

  • 1.Writing good research papers
  • 2.hello! I am Kai Gao I am here to share my experience as a PhD student/candidate kaigao@scu.edu.cn 15120003793 2
  • 3.My PhD Career Started my PhD program Trying to find a research topic After 3-4 rejections, my first paper was accepted by IWQoS on the very day of April 15th, 2017 The paper was later extended and was accepted by TON in 2019 2013.9 2012.9 2015.9 2014.9 2016.9 Published the first paper as a collaborator Assigned a problem and submitted my first paper (INFOCOM) but it was rejected 2012-2018 PhD Candidate at Tsinghua University Passed the Qualification and became a PhD candidate Yes, with one-year extension 2016-2017 CSC Exchange Program at Yale University 2017.9 Participated in the CSC exchange program 3 2018.9 Two papers were submitted to INFOCOM 2018 and one got in Two papers were submitted to SIGCOMM 2018 and one got in Collaborated in 3-4 papers accepted by ICNP and JSAC
  • 4.Skills Find a problem Propose a solution Technical contents are the foundation of a good research. But good writing helps to ○ ○ better convey your ideas and highlight the novelties increase the chance of being accepted 4 Present the work ○ research paper ○ talk
  • 5.In this talk, you are about to hear ○ Tips to improve the writing ✓ They work for me, and they should work for you too ○ Tools to help with the writing ✓ Formatting, citation, and collaboration 5
  • 6.1. Tips for High-quality Academic Writing How to better present your research 6
  • 7.Tip #1: Learn from the masters A good paper is often well-structured. If you don’t know how to structure your paper, find a good example (Propane, SIGCOMM’16 best paper) Find a good paper as your template ○ ○ ○ preferably from the same conference/journal similar to your paper in terms of methodology (system paper, theory paper, etc.) follow their structure and content organization An example for writing system papers especially with new interfaces 7
  • 8.Tip #2: Use the top-down style Always let reviewers know what they are about to read beforehand. (Trident, SIGCOMM’18) Make it clear ○ ○ what this section/paragraph/ algorithm is about which part is the most important and why the reviewer should pay attention to it 8
  • 9.Tip #3: Use figures wisely A figure is to make things simpler to understand not to make it complex (KeySight, ICNP’18) (ONV, TON’19) A very clear illustration of the algorithm A very clear illustration of the design space 9
  • 10.Tip #4: Highlight your insights We are computer scientists, not archemists. A good insight is a sign of “scientific” understanding of the problem. (RDD/Spark, NSDI’12) Explain why your solution solves your problemwell:○ ○ ○ specific properties no one used before specific patterns no on observed before … (your insights) After reading this paragraph, it gives a good sense why RDD works 10
  • 11.Tip #5: Be your own reviewer Your writing only improves when you know it can be improved. Also, it’s better for you to find a problem before reviewers do. ○ ○ TRUESTORY:I rewrote a whole section 3 hours before the IWQoS submission deadline. Start writing early and run multiple “write-review” iterations. It is usually better to review your draft after a few hours It’s never too late to make a change so don’t easily get settled My collaborator and I rewrote two sections in our INFOCOM submission after the deadline was extended for 8 hours. Both papers were substantially improved. 11
  • 12.2. Tools for Efficient and Enjoyable Academic Writing How to make research & writing easier 12
  • 13.Use LaTeX When you should use LaTeX ✓ ✓ ✓ When you should not use LaTeX easier typesetting ○ math, source code, etc. ○ resubmission to another venue easier for collaboration ... ✘ 13 if your collaborators don’t use it
  • 14.Use a Good Reference Manager What makes a good reference manager (I use Zotero but I don’t recommend it) you must be able to ✓ ✓ ✓ search related papers categorize related papers and take notes access references directly in your favorite editor 14
  • 15.Use a Good Editor What makes a good editor ✓ ✓ ✓ (I use Spacemacs a lot and I recommend it) You feel comfortable typing in Good integration with your reference manager Simplify collaboration 15
  • 16.Use a Good Editor What makes a good editor ✓ ✓ ✓ (I use Overleaf sometimes and I recommend it) You feel comfortable typing in Good integration with your reference manager Simplify collaboration 16
  • 17.My Writing Gear ○ ○ ○ ○ ○ LaTeX (texlive + latexmk, and a bunch of self-made packages) Git (github & overleaf) Spacemacs (with the latex layer) Zotero (with the Better Bibtex extension) Google scholar 17
  • 18.Takeaway WritingTIPS:1. 2. 3. 4. 5. WritingTools:Learn from the masters Use the top-down writing style Use figures wisely Highlight your insights Be your own reviewer 1. 2. 3. 18 LaTeX Reference manager Editor (spacemacs & overleaf)
  • 19.thanks! Any questions? kaigao@scu.edu.cn 15120003793 19