The beginner’s checklist for avoiding writing blunders – I

You have done a great piece of research; designed the perfect experiment or built the perfect learning model, the results are promising and it is time to submit a paper. But wait…, only two days left before the conference deadline and you find it is pretty boring/tiresome to write things. Sounds familiar?

Many early research students do not realise the importance of the writing element in research until they start writing their first paper or worst until they see their first set of reviews. In some cases, I am surprised by the amount of typos and basic writing errors I see in papers I review. In this post I try to provide a checklist for early research students to ensure that their first serious research submission, be it a master’s thesis, conference paper, or a journal paper, is free of writing blunders. By looking at the list, one could argue that these are very obvious and trivial errors. Yet, you will be surprised to see how frequent they are, especially among the writings of early stage research students.

This is possibly the first of a series of posts, in which I plan to cover several academic writing related topics.

1) Check spelling mistakes: Use the in-built spell checker of your text editor to check spelling errors. Grammatical errors are hard to catch. I will discuss this in a seperate post.

2) Check the incomplete citations (??): Incomplete citations can happen if you keep a citation placeholder (e.g. \cite{} ) but forget to fill it or add some new citations to the bib file but forget to compile it. Once you generate the PDF do a search for ??s to avoid this ugly mistake. Same can happen with the latex \ref{} command for figures and tables.

3) Figures and tables: It is good to have figures and tables in research papers. They summarize results and can provide a quick overview to the readers.

  • Make sure all the figures and tables are referred in the text (In the order they appear)
  • All tables and figures must have captions
  • Figures axes must have titles and the units have to be clearly mentioned

4) Space errors: Quick writing can insert space errors. Many of them can’t be captured by searching in PDF readers and requires careful proofreading.

  • Extra space at the end of a sentence (Example: The simulation results showed that our hypothesis was correct .)
  • No space when starting a new sentence. (Example: The simulation results showed that our hypothesis was correct.However, …..)
  • No space before a reference (Example: Smith et al.[10] showed that a deep model ….)

5) Correct use of commonly used abbreviations/notations/symbols

  • Smith et al. [10] (et al. with a period is the correct usage)
  • e.g. is the correct usage of “for example” (similarly i.e.)
  • If a sentence ends with a footnote symbol, the period comes before the footnote not after. Same for comma as well. (Example: Alexa lists the top-100 websites in the internet.1

AI & Machine Learning for Cybersecurity – A Compilation of Resources

Last few weeks, I was preparing some lectures on AI & Machine Learning Methods for Cybersecurity. While I have been working in the area, I haven’t taught such content in the past and this was a whole new experience for me as well.  Surprisingly, I did not find much resources that provide a structured view on the subject. Below is a set of resources I found highly useful whilst preparing my lectures.

1) CS 259D Data Mining for Cyber Security, Stanford University – Autumn 2014
(course web page)

This probably is the only university course I found which is doing exactly I was preparing for. It  covers a range of topics such as behavioural biometrics, deep packet inspection, and phishing detection. While it appears that the course is not offered after 2014, the topics it has covered in the last offering is comprehensive and still highly relevant. One topic I might add to this already comprehensive list of topics is probably adversarial machine learning (which was probably has not become mainstream by then).

2) Machine Learning and Security: Protecting Systems with Data and Algorithms By Clarence Chio & David Freeman
(link  to the page of the book)

I enjoyed reading this book as well as trying out various example codes released with the book as Jupyter notebooks that can be found here.  Major part of the book follows the format in which different machine learning concepts that are related to security are introduced with worked examples. Some interesting chapters include network traffic analytics, adversarial machine learning, and malware analysis.

3) Research paper compilations

There were several repositories that are collections of papers related to the use of machine learning in security applications.

  1. The Definitive Security Data Science and Machine Learning Guide:  In addition to collection of papers, this website provides a compilation of blogs, datasets, books, and presentations on the subject.
  2. Machine-Learning-for-Cyber-Security
  3. Collection of Deep Learning Cyber Security Research Papers.