Effective tools for computer systems research
Note:Reading [2020-06-11 Thu 11:40]
Before diving into a paper, know what you want to get out of it.
Ideally, an abstract reveals the problem that the paper tries to solve, explains how it solves the problem, and what the results are. A well-written abstract provides enough information for you to decide if the paper is worth diving into.
Adopt the mindset that you are reviewing rather than reading a paper.
Your null hypothesis should be to distrust a paper’s results and only through sound reasoning and rigorous methods can a paper change your mind.
Things to think about
Briefly summarise the paper’s method.
– Is the method section detailed enough to facilitate reproduction?
What are the paper’s assumptions?
– How robust and realistic are these assumptions? – Can you think of counter-examples to these assumptions?
What are the paper’s key results?
– Do these results contradict or confirm existing work?
How could the paper (its method, presentation, or results) be improved?
What follow-up research questions come to mind?
What’s the paper’s conclusion?
– Do the results support the conclusions
- No necessity to read a paper from cover to cover, I guess, unless specifically insisted by Profs :O
- (Trial run in lab?) Maybe a reading group, to increase reading papers from specific areas
Maintaining a Bibtex file for research, currently in progress, let’s see how it turns out.
paraphrasingto read through the introduction,conclusion to get a better idea, then see if it makes, focus on review than reading a textbook, accept the fact that the author can be wrong, see if possible to replicate the results, keep checking Scholar alerts/Arxiv not sure about that, well scihub and mailing the author for work :)