Notes/Tutorials on best practices in research
Resources for CS Research
- [Google Scholar] (https://scholar.google.com/) * [CiteSeer] (http://citeseerx.ist.psu.edu/index;jsessionid=FBBDBD6285DBEEE56F7D15668C79C657) * [ACM Digital Library] (http://dl.acm.org/) * [IEEE Xplore] (http://ieeexplore.ieee.org/Xplore/home.jsp) * [DBLP] (http://dblp.uni-trier.de/) * Last but not least, don't forget your local library Be wary, these platforms are often highly biased towards recent results.
- Survey Papers: Good to accumulate relevant citations (also tutorial). * Vision/Position Papers: Understanding what questions are important for community (not result oriented). * Experimental Papers: Fundamental paper accumulating experimental results. * Theoretical Papers: Fundamental paper accumulating theoretical results. Little known fact: If you'd like to use a paper that you can’t find for free online, you can always email the author.
- [Mendeley] (https://www.mendeley.com/) * [Papers] (http://www.papersapp.com/) * [EndNote] (http://endnote.com/) * [BibDesk] (http://bibdesk.sourceforge.net/) Organization is necessary! Read 100+ papers/year to keep up with industry discourse.
- Be suspicious of papers that claim to be the next best thing and don’t admit any limitations - Reference limitations or future work of other papers for interesting material to research - Try to focus on good science as opposed to research fads (cough, cough, deep learning) - There are many asked questions that we are now developing the tech for, so research past papers - Use insights of references, but avoid duplication; use methodology - Find earliest known results on a topic - Add a few irrelevant papers if they look interesting, could help to improve research question - Theorize how those in other fields would go about your question - Specify your research question, if you’ve found it’s already been answered, write a follow-up question (It’s a good thing if you find a paper that’s already been tackled - it means you’re asking the right questions)