Academic Research in Simple Steps

Through last two years, I finished my master of science in cairo university , and started PhD of science at Deakin university, and although I made good enough master of science with good publications (thanks to my supervisor), I have to state that the research process is not clear enough in Egypt, which leads many researchers to make the same mistakes and waste lots of time in already known things.

The research process is straightforward, but unfortunately it is not clear enough in Egypt, and many of the researchers are taking it as kind of certificate for social prestige rather adding real value to the science.

To make a good research, the researcher has to :

1. State the motivation of the research that is either facing a problem in your work or life and try to fix it (researches of car safety or cancer cure) , or to have innovative idea that u think it will add new value to the life like (e.g.nano technology).
2. Identify the problem statement (what is the problem u want to tackle and why)
3. State your objective (what do u want to achieve from this research? )
4. Build strong literature review, that is lots of readings of best papers related to your problem, summarizing them, criticizing them and comparing techniques and results in between them.
   
   Best papers can be identified as :
   
   • earlier than 4 years ago: so it is not mentioning obsolete techniques
   • it is earlier than 4 years it should be highly cited by many other papers (10 citations at least)
   • be published in highly ranked journals or conferences. In NLP u can find ACL, EMNLP, COLING, EACL, AAAI
   • Otherwise, it is nice to read other papers as u may find some good ideas in them but give them lower priority in reading, and take care that some papers put fake results (from low ranked conferences) , and some papers are useless (not cited).
5. the new possible solutions to your problem and the possible scientific contribution which will be the added value to the science
6. your solutions practically , and it would be better to do it using standard and open-source tools, so evaluation committees can evaluate it easily , and further researchers can extend your research without re inventing the wheel.
7. your stated solution on your experiments using standard datasets and evaluate your results using standard evaluation techniques (benchmarks) and finally compare your results to other people's work

N.B. keep documenting and taking notes of every action you are doing in your research from your first day, as you will need them all again in writing your papers and thesis

N.B; To make a master that is good enough, u have to publish at least two international papers. and to make good enough PhD you have to publish at least 5 international papers

For publications: Whenever you want to publish your paper try always to target conferences and journals that are level A or level B because any publication in any journal or conference that is ranked less than level B is usually weak publication , and any publications less than level C are not even considered as publication in any reputable research lab.

I hope these words will help any new researcher anyhow.

These are my limited knowledge and recent experiences, and I hope any more experience researcher share his experience with me


Ahmad Hossny
qasaqis.blogspot.com