Scientific and Academic Research: A Survival Guide
1. Scientific and Academic Research:
A Survival Guide
Centre for Vision, Speech and Signal Processing (CVSSP)
Electrical and Electronic Engineering Department
University of Surrey
3. Why do we do research?
− Explore new ideas and explore new horizons.
− Develop novel solutions and prove they work!
− Solve unsolved problems and/or find new problems.
− Publish papers (and travel to conferences in nice places)
− Attract funding
− Create new systems/solutions (demonstrators) - to get more funding
− Enhance our CV
− Secure an academic/industry job
− Or maybe impress people (friends and family?)
4. Why do we do research?
“One of the great things about science is
that it can change the way we see the
Shankar Vedantam, Hidden Brain, NPR
5. Questions to ask before (and while)
These questions can be applied to publishing a new
paper, starting a new project or applying for a new
6. 1. What is the problem?
−What problem you are trying solve.
−Not in vague and technical terms but in a simple
−Pause for a minute and think what problem(s)
exactly your work is trying to solve.
−Designing x or building y is not solving a problem;
what do that x or y will do to overcome some
issues/challenges that we are facing today?
7. 2. Why is this problem important?
− Most people will say the problem(s) that they solve are very
− Here you need some clear statements and evidence to show
that the problem is an important one.
− Working on a complex problem does not necessarily make it
important, nor does linking it to a complex subject (but this
will definitely make it harder and it will have a higher risk of
− Focus on the problem not the solution (at least, to begin
8. 3. Who will benefit from this research?
− Again you need to be specific and say who, what sectors, why
and how these will benefit from your research.
− This can be an applied or fundamental research; both are
(hopefully) driven by curiosity; however:
− Applied research: This can be linked to a popular (or
unpopular) problem/issue related to economy, society and
other technical and none-technical challenges that (hopefully)
will benefit people and the cost (time, complexity,
investment) of doing this will be worth the benefits.
− Fundamental research: If you do pure science, you may argue
how this will change the way we see (or perceive) the world.
9. 4. How is this different from what has already been
− You need to compare and contrast your work with some of
the existing and well-known solutions.
− If there are several papers/solutions/projects in the area of
your work, you need to have some clear understanding of
these and provide statements that stand out quickly; you
should probably avoid generic statements.
− You obviously need to support your claims with some
10. 5. What are (or will be) the key novelties?
− Scientific research advancements.
− Industry/use-case solutions that can benefit from this work
(you need to explain what will be novel about them).
− You can discuss how you have changed (or will change, if you
are writing a proposal) the way that people look at a problem
or view a domain.
− You may propose a completely new
product/service/algorithm/technology/… or something that
no one has done and/or thought about before - but to make
it interesting from a research point of view you need to say
why it is new or better, compared to what is already out
11. 6. Why is this a challenging research and development?
− You need to clearly explain why simple or existing solutions
won’t work or are not sufficient.
− e.g. centralised solutions are not suitable for large scale …
because they have large overhead, so we propose …. .
− Justify why one should fund/support this research for x period
with n number of people; or, if you are trying to publish your
work, justify why this work has been so interesting and
fascinating (painful?) that it deserves to be published.
− Put simply, you need to show why an easy fix wouldn’t have
12. 7. What is the main impact(s) of this work?
− Specific statements on how this will contribute to … and the
well-being of people, society, safety, etc… and why this
impact will be (or is) achieved by this work/project and not
− Otherwise everyone can say we have an impact on x, y, z,; it
is very important to say why (and how).
− Think of your work’s unique offerings.
− Put yourself in your audience’s/customers’ shoes.
13. 8. What will you require to do this work?
− More funding, more help, …
− You need to justify the resources that you are asking for and
show why your work is worth it.
− For proposals, you also need to explain why you believe the
team that you have/propose is the best combination of
people/expertise to deliver this.
14. 9. What are the tangible outcomes?
− What your results will look like.
− What products and services you will offer by end of this
project (at what TRL levels).
− Discuss your plans to use/dissaminate the results.
− If this is a paper:
−Add links to underlying data/code (if there are no
−Discuss the reproducibility of your results.
15. 10. Who else could be involved?
− Stakeholders, user groups,
− Other teams, other disciplines, your peers.
− Think of a wider audience for your research.
19. “Anyone who has never made a mistake
has never tried anything new.”
20. Fault tolerance in research
−Do not compromise the quality and results,
−However, be prepared for failures,
−and have a recovery plan.
21. “Fear is the main source of superstition, and one
of the main sources of cruelty. To conquer fear is
the beginning of wisdom”.
22. “Failure is so important. We speak about
success all the time. It is the ability to resist
failure or use failure that often leads to
greater success. I’ve met people who don’t
want to try for fear of failing.”
− It is not equivalent to complexity
− or obscurity
− It’s the art of seeing things in a way that hasn’t been seen
− Finding metaphors and explaining things using the new
− Looking and looking until you unlock the problem.
31. “The problem is never how to get new, innovative
thoughts into your mind, but how to get old ones
out. Every mind is a building filled with archaic
furniture. Clean out a corner of your mind and
creativity will instantly fill in”.
32. “The world is full of obvious things which
nobody by any chance ever observes.”
The Hound of the Baskervilles
36. Technique vs. Problem Orientation
− “no technical skill is worth more than knowing how to select an
existing research project”.
Peter J. Feibelman, A PhD is not enough
− Focus on the problem; you can always learn the techniques.
− And also know/learn the basics and principal concepts and
theories in your field.
38. On Problem Solving
−“When you follow two separate chains of thought,
Watson, you will find some point of intersection
which should approximate to the truth.” –
Sherlock Holmes, The Disappearance of Lady Frances Carfax
−“Always approach a case with an absolutely blank
mind. It is always an advantage. Form no theories,
just simply observe and draw inferences from your
Sherlock Holmes, The Adventure of the Cardboard Box
− Scientific process relies on organised scepticism.
− Your ideas and your work should be clear and reproducible.
− Your work and ideas will become really useful if other people
agree with them and use them.
− Write/present them in accessible forms.
− Make your underlying assumptions clear.
− Make your data/results/code (depending on the IP/privacy/sensitivity)
− Respond to queries and questions.
43. Generosity in research
−Be generous and make friends.
−IP and protection are important.
−But, research and innovation also come with sharing
knowledge and discussing your work with your
−This will always keep you one step ahead…
44. Legacy of your research
“If in 100 years I am only known as the
man who invented Sherlock Holmes then I
will have considered my life a failure”.
Arthur Conan Doyle
45. Institutionalised Conflict
− People (read managers) [may] make mistakes; sometimes
these mistakes [may] hurt other people.
− Competition can cause conflict.
− Differences of opinion, and different methods and approaches
to solving/addressing the problem can cause conflict.
− “Successful collaboration is possible when one or both contributors
have established reputations, or when each researcher brings a
different identifiable skill to the collaborative project”.
46. Find a good mentor
− Someone who doesn’t see you as a competitor.
− Don’t be shy about getting to know people outside your
− Ask for help; everybody likes to give advice
− A good mentor can be a former advisor, a senior member in
your institution, someone you meet at a conference
(that’s why you should be well prepared for your
conference/workshop talks and attend important talks at
conferences and ask questions).
− Never send angry emails…
− Never send emails while you are angry…
− Always wait and don’t send emails if you cannot think
− Sometimes if you wait enough an annoying matter (which at
the time may seem to require immediate attention) will solve
itself or will get solved; just be patient. Avoid the impulsive
− Don’t threaten to leave your job (unless you already have a
very good offer) and never leave (without having a better
offer that is confirmed) just because you don’t like … or ….
− Even if you decide to leave a place, always leave on good
− If you are upset find the right time (and the right person) and
express your frustration but resist the temptation to talk
about someone who is not present.
− You need to set your own targets; who you want to be,
where you want to be.
− There is a lot that you are capable of doing and there are lots
of things that you may be asked to do (or volunteer for); but
always think about their importance and your obligations.
− Don’t overdo it, and don’t avoid helping others - have some
− Learn to say “No” – thank you for your invitation. I appreciate
your interest in our work, but unfortunately I am overcommitted at
− If you are involved in a collaborative
project/meeting/discussion/proposal that seems to be going
nowhere, politely (and slowly) reduce your involvement.
− You are in academia so there must always be some meetings,
teaching obligations, tutorials, lab supervision that [can]
overlap with the (group in question's) meeting.
− Before saying “Yes” think: 1) do I have time for this? 2) is this
really important? 3) Will I or someone else will benefit from
52. Work with your peers
− Help them because one day you may need their help.
− Try to find a group of people with whom you would
genuinely like to work and collaborate over time; from your
discipline and other disciplines as well.
− But, have clear targets; if you only meet and talk and never
get anything tangible done, you will get tired and others may
get tired of you (the latter is more dangerous).
53. Choosing an advisor
− A prominent scientist versus a young advisor.
− Look at their track record; find a group to whom you can
offer something valuable and from whom you can get
sufficient advice and mentorship.
− If you are applying for a postdoc, always consider whether the
postdoc role in the place to which you are applying will help
you to land a permanent job in the field/sector in which you
54. Key to success as a postdoc
− Finish something.
− Make yourself known and useful to your colleagues and your
− Publish and make your results available/visible to your future
− You need to develop/create something interesting that you
can talk about.
− Don’t be a slave to your postdoc advisor.
Peter J. Feibelman, A PhD is not Enough.
Naomi Oreskes at TED
57. Giving talks
− Know your work.
− Be well prepared.
− Respect the time and focus on the key points.
− Avoid self-promotion; your work should speak for itself.
− Understand and respect the needs/interests of your audience.
− Your talk is for your audience not for you; so think and
investigate what it will be important for your audience to
learn/know. Put yourself in their shoes and think what you
would expect/want to see or hear.
58. Giving talks
− Talks for job interviews, your colleagues and collaborators,
conference/workshops and public talks.
− Use your time efficiently and do not overestimate or
underestimate your audience.
− Provide some basic information about your work.
− Tell a good story; something that people will follow and
59. Giving talks
− Often you don’t need to show an agenda.
− If you are showing graphs, at least say why you are showing
them and highlight the important points.
− Show mathematical equations and complex formulas only and
only if you don’t have anything else interesting to show
(unless maths is your field of work).
− The opening to your talk is important.
60. Giving Talks
− Practice and practice.
− Be prepared.
− Slides should look professional.
− If it is for a job interview, learn about your audience, their
work, interests; link your work to some of their projects and
work. [Indirectly] tell them why you think your expertise will
be valuable and will complement their work.
− Don’t oversell/over-promote your work but at the same time
don’t be too humble.
− Be enthusiastic about your work.
61. Use humour (if you can) but don’t be offensive
Image source: Stocky, NPR
62. Giving talks – Q&A
− Appreciate the audience’s time and attention.
− Don’t be too defensive (or aggressive).
− Listen to the questions and answer them to the best of your
− If you don’t know the answer, give a response based on your
analysis but say that you don’t know the exact answer to the
− Some people may come across as aggressive or ask their
question in a way that makes you uncomfortable; the
audience will hear and feel this so just be cool and try to
answer the question with logical statements and according to
scientific facts and evidence.
63. A Career in Academia
− Establish a reputation for your work (different from trying to
become famous- this can come later).
− Make friends and show that you are willing to work with
− Study the department(s)/group(s) that you would like to join
and see what are their key criteria for hiring academic
− Have a web presence.
− Be a good citizen of your field/community (but don’t overdo
− Show interest in teaching and try to gain some experience
(TA, Tutorials, talks).
64. A Career in Academia
− Create a vision for your future research and align it with the
priorities of funding bodies (and talk about this in your
interview and/or whenever suitable with your Manager, HoD,
Dean, Provost for research, …).
− Be prepared to attract research/industry funding.
− Interact with people- sometimes calling someone on the
phone or offering to go and meet them personally can turn
out to be really, really, helpful.
− Stay out of office politics if you want to do something
65. A Research Career in Industry
− You can probably get a permanent job (if there is anything
like this anymore) more quickly.
− Your job description will probably be simpler than your
− You will work in a “managed” environment.
− Your research will most probably be around your company’s
field of work and interests.
− Depending on the company’s interest, you may be asked to
generate patents, applications, demos, or papers. Try to find
out about this before you join and see if that’s what you are
“Words are, in my not-so-humble
opinion, our most inexhaustible source of
J. K. Rowling, Harry Potter and the Philosopher’s Stone.
−They will stay in libraries and most probably be
accessible for ever.
−So think about this when you submit them.
−This shouldn’t discourage you from sending an early
stage work; but be precise and clear.
−Remember, your students and colleagues may read
your work in the future; or you may even read it
yourself one day.
−Choose the right journal/conference.
−You need to publish a certain number of papers but
quality is more important.
−People spend their time reviewing/reading your
papers; respect their time and effort and write
something pleasant to read, that the readers will
learn from and will find interesting.
−Survey papers are not a summary of the published
work in your domain.
−The literature review/related work section of your
paper is not a set of sentences about some related
work with a citation.
−Avoid any “magic” in the the results section- make it
clear how you have obtained your results and under
−Talk about the limitations of your work and link
these limitations to the direction of your future
−In your discussion/conclusions, if you can, talk about
the lessons that you have learned (and mistakes that
you made during the work).
−Make sure your claims are reproducible.
−Acknowledge other people’s contributions.
71. Survival Checklist
− Consider your audience - their interests, priorities, …
− Be a resourceful colleague/collaborator but also learn to say
− Define your priorities.
− Meet and network with people within and also outside your
− Create a plan for your publications and target key events.
− Apply for funding (consider the 10 questions that we
− Remember and remind yourself of your goals.
− Several people have contributed to the thoughts and
discussions related to this talk, including Prof Amit Sheth, Atti
Emecz, Prof Hamid Aghvami of King’s College London, my
students: Yasmin Fathy, Daniel Puschmann, Nikos
Papachristou and many other friends and colleagues.
− Some parts of the slides are adapted from Peter J. Feibelman’s
“A PhD is Not Enough” (one of my favourite books), Ray
Kurzweil’s “How to Create a Mind”, NPR’s Shankar
Vedantam and his “Hidden Brian” Program, NPR’s Guy Raz
and his “TED Radio Hour” program and several brilliant
speakers at TED.
Thank you for your attention.
75. Links and sources
− Twenty things I wish I’d known when I started my PhD
− Done is better than perfect: overcoming PhD perfectionism
− “Thinking Fast and Slow”, Daniel Kahnemann.