Andre Geim on writing

 Andre Geim describe what he learnt from his Russian literature teacher (from the biography)

"I also learned an important lesson from my tutorials in Russian literature. My tutor said that what I was writing was good but it was clear from my essays that I tried to recall and repeat the thoughts of famous writers and literature critics, not trusting my own judgement, afraid that my own thoughts were not interesting, important or correct enough. Her advice was to try and explain my own opinions and ideas and to use those authoritative phrases only occasionally, to support and strengthen my writing. This simple advice was crucial for me – it changed the way I wrote. Years later I noticed that I was better at explaining my thoughts in writing than my fellow students."

Geoffrey E. Hinton's advice for students (from Coursera)

This part is taken from an Interview of Prof. Geoffrey E. Hinton by Prof. Andrew (in Coursera course)

What is your advice to students to want to pursue a career in Deep Learning?

(I thin that the answer given by Prof. Geoffrey E. Hinton is useful to any researcher irrespective of the field of research)

Here is the answer.

"Read the literature, but don't read too much of it. So this is advice I got from my advisor, which is very unlike what most people say. Most people say you should spend several years reading the literature and then you should start working on your own ideas. And that may be true for some researchers, but for creative researchers I think what you want to do is read a little bit of the literature. And notice something that you think everybody is doing wrong, I'm contrary in that sense. You look at it and it just doesn't feel right. And then figure out how to do it right. And then when people tell you, that's no good, just keep at it. And I have a very good principle for helping people keep at it, which is either your intuitions are good or they're not. If your intuitions are good, you should follow them and you'll eventually be successful. If your intuitions are not good, it doesn't matter what you do."

[Prof. Andrew Ng: I usually advise people to not just read, but replicate published papers. And maybe that puts a natural limiter on how many you could do, because replicating results is pretty time consuming.]

"Yes, it's true that when you're trying to replicate a published you discover all over little tricks necessary to make it work. The other advice I have is, never stop programming. Because if you give a student something to do, if they're botching, they'll come back and say, it didn't work. And the reason it didn't work would be some little decision they made, that they didn't realize is crucial. And if you give it to a good student, like for example. You can give him anything and he'll come back and say, it worked. I remember doing this once, and I said, but wait a minute. Since we last talked, I realized it couldn't possibly work for the following reason. And said, yeah, I realized that right away, so I assumed you didn't mean that. "

[Prof. Andrew Ng: ny other advice for people that want to break into AI and deep learning?]

"Basically, read enough so you start developing intuitions. And then, trust your intuitions and go for it, don't be too worried if everybody else says it's nonsense."

"If you think it's a really good idea, and other people tell you it's complete nonsense, then you know you're really on to something."

I think that this advice is applicable to any technical field. Some researchers spend too much time on reading and this lead to too much bias in the published results. If researchers read just enough so that they can think and work on a problem, they can contribute something new in that field. Otherwise, they will be doing just what others have done (mostly incremental work).

Another thing I understand is that students and researchers should try to replicate seminar research papers from the scratch. This help the researchers understand the process the top researchers have undertone. This often provide the skills needed to find out something new in that field and contribute novel works in that field.

Reference:

Weblink: https://www.coursera.org/learn/neural-networks-deep-learning/lecture/dcm5r/geoffrey-hinton-interview

Symbols above the letters (for scientific publications)

When writing papers, we often come up with author(s) names with symbols above some characters of author(s) names.

How do you use those symbols in LaTeX?

Here is a Wikipedia link where the details are provided.

https://en.wikipedia.org/wiki/Diacritic

In order to use in LaTeX, a number of answers in the following SE site is given.

https://tex.stackexchange.com/tags/accents/info

Some of the frequently used symbols are (from SE site):

Text mode

Plain TeX makes it possible to typeset the most commonly used accents:

• \` (grave accent): à
• \' (acute accent): á
• \^ (circumflex or “hat”): â
• \" (umlaut or dieresis): ä
• \~ (tilde or “squiggle”): ã
• \= (macron or “bar”): ā
• \. (dot accent): ȧ
• \u (breve accent): ă
• \v (háček or “check”): ǎ
• \H (long Hungarian umlaut): ő
• \t (tie-after accent): a͡
• \c (cedilla): ş
• \d (dot-under accent): ạ
• \b (bar-under accent): ο̩
• \k (ogonek): ą

The Unicode character encoding UTF8 includes several special characters and characters with accents. The following code specifies that the encoding of the LaTeX document source file is UTF8. As font encoding is specified T1, because it supports the encoding of extended character sets in fonts:

\usepackage[utf8]{inputenc}
\usepackage[T1]{fontenc}

Of course, the encoding in the text editor needs to be set to utf8, as well.

Math mode

The following commands may be used only in math mode to produce accents;

• \hat{o} (circumflex): 
• \widehat{oo} (wide version of \hat over several letters): 
• \check{o} (vee or check): 
• \tilde{o} (tilde) 
• \widetilde{oo} (wide tilde) 
• \acute{o} (acute accent): 
• \grave{o} (grave accent): 
• \dot{o} (dot over the letter): 
• \ddot{o} (two dots over the letter): 
• \breve{o} (breve): 
• \bar{o} (macron): 
• \vec{o} (vector (arrow) over the letter): 

What is your biggest lesson after publishing in Nature Journal?

"What is your biggest lesson?" This is the question that Nature asked researchers published in Nature recently.

Christina Hicks:
"I need to focus on myself for me. It’s easy to become really busy. I run up and down the mountains in the UK Lake District for fun. My brain gets cluttered and overworked if I don’t take a break. Running helps me to slow down."

Johan van den Hoogen:
"Working on that nematode paper helped me to appreciate the value of outsourcing — I didn’t develop the models in it. You should let other people do the things that they’re good at. I understand the models in the paper, but it would have taken me a year and half to create them on my own."

"My biggest realization is that you don’t need to move up the academic career ladder to have a satisfying career in science. The moment I stopped worrying about advancing in academia marked a change for me."

Stephanie Ellis:
"You have to have people around you who can point out the weaknesses in anything you do."

Oscar Serrano:
"Some relationships are more mutually beneficial than others. You can collaborate for years with someone and then realize it’s been a one-way street. You’re sharing ideas and resources with them but getting little in return. But other people really do give back as much as they get and really help you grow your career. I want to be the person that people want to collaborate with because it’s reciprocal."

Here is the full article from Nature.com

How to write a paper quickly?

Top 10 rules

Note that writing a paper to journal is a painstaking effort and takes a lot of time for researchers in their early career (some times even for experts too). Sometimes, it may be possible for an expert to write a complete paper from conception to submission format. These are considered to be an exception. Personally, I take at least six months to come with the first draft of the manuscript. Here is one of the ways to write papers quickly for a journal of conference. 

1. Start with the hypothesis and start writing. Don't procrastinate. Write from the starting of the work, when you are doing experiment, write the methodology.
2. Review the notes and renew the literature search, and refine your conception and hypothesis
3. Determine who your audience is - most likely reviewers, so get their concerns firstly addressed
4. Create the outline and get the big picture done, i.e., complete the first draft while resisting the temptation to correct and edit as you go; the logical sequence of data/tables/figures may be the outline; while writing the first draft, take notes indicating what references might be needed and would be about, but don't stop to collect the references
5. Begin with the easier part of the task - experimental section rather then the introduction; then follows the results and discussion section
6. Then comes the really hard part - critical editing where you make sure that the English is coherent and the science is correct
7. Write the conclusion in a numbered format
8. Then comes the abstract and the acknowledgements
9. Now comes the introduction, the two most important things to address in which are the purpose and relevant background
10. Then collect the references.
11. Use Latex for writing your paper. No matter how complicated your paper is, LaTeX will give you a best final draft.

Reference

https://www.youtube.com/watch?v=UY7sVKJPTMA

Writing Good Papers (Prof. Michal Lipson)

Top advice form Prof. Michal Lipson, 

1. Don't write paper for those 10 people who is specialized in your area (who will read no matter what). Our goal should be to reach much broader community (broader audience) 
2. In one minute of grasping you paper, the reader should know what you did
3. Paper should be crisp clear and easy to read (should not be hard to read). Writing your research without clarity may have no effect even though your work has potential to form a new field in research
4. Tell your story in one minute.

Golden Rules

1. Cut! This seems to be the most important rule. Reduce any unnecessary word or phrases (cut many, anything subjective). Cut efficiently. Say aloud and write it. When we say something, we say clearly and simply. Just write as you explain to some one else. After writing, hunt down every word that will distract the reader from the main point you make. Very, really, basically, quite, generally - these words don't add much. Write without these sentences. Avoid starting with "there are". Following are the unnecessary phrases: for the most part, for the purpose of, in the case of, in the final analysis, in the event that, it has been estimated that, it may be argued that (come straight), as far as we know, as we know. Remove clunky phrases (with the word in bracket):  A majority of (most), A number of (many), Are of the same opinion (agree), At the present moment (now), By means of (by), Less frequently occurring (rare). Wordy vs Pointed: in spite of the fact that (although), in the event that (if), period of four days (four days), refer back (refer), shorter/longer in length (shorter/longer), had been previously found (hand been found). Ask these questions again and again: Is this word/phrase necessary?, What happens if I take it out? If it is not critical (not changing anything), remove it.

2. Use active voice. When use passive voice, the sentence becomes hard to read, and not actually gives who did the experiment. Use we found major differences.... We did this, We did that, etc.

3. Use parallel construction: If you want to be a good researcher, you must study hard, listen well, and think critically about the literature.

4. Start paragraph with main point: Start the paragraphs with your main point then expand (Professor's favorite). This is important to write a crisp clear paper. Every paragraph should start with the main idea of the whole paragraph. Rest of the sentences are should be just details explaining the first sentence. If you take first sentences of all the paragraphs and put them together, it should form the abstract.

If you want to write a sentence and if you don't know how to order it, always bring the important points front.

Know the difference between the introduction and discussion. 

In introduction, start with what is available, then explain what is missing, then your work (i.e., the introduction moves from general to specific)

In the discussion, at the end, say what amazing thing you did is going to revolutionize the word  (The discussion moves from specific to general). You know the simplification, why and what you did some body else would care, and why the deficiencies you have might or might not be fundamental (which is pretty important). Put your work in the big picture. Your performance is not the best. But, put the minor drawback of the work. 

In results, you show the number. In discussion, you show the big picture. The discussion should be the big picture. Not the little details.

Most important things not to put in the discussion: 

Do not simply repeat what is in the Results

Do not try to explain every minor flaw

Do not try to explain away every unexpected result

Do not exaggerate or make extravagant claims 

Don't hedge

(Editors and reviewers may read only this part, if they are busy. They don't care about the number. They care about the big picture.)

Need to attract very wide audience. But don't over claim and don't exaggerate.