How To Give A Great Research Talk

How to give a
great research
talk Peyton Jones
Simon
Microsoft Research
Cambridge
https://www.microsoft.com/en-us/research/academic-program/give-great-researc
h-talk/
Photo © James Millar/TEDxExeter

Why you Research is communication
should • Think: how often have you said

“I’m really glad I went to that
listen to talk”
this talk • Some simple, actionable ideas
that can make your talks much
better
• You will have more fun
• A research talk gives you
access to the world’s most
priceless commodity: the time
and attention of other people.
Purpose
The • To impress your audience with
your brainpower
purpose • To tell them everything you
of your know about your topic
• To present all the technical
talk is details
not:
The • To give your audience an
intuitive feel for your idea
purpose • To make them foam at the
of your mouth with eagerness to read
your paper
talk is: • To engage, excite, provoke
them
• To make them glad they came
Audience
Your The audience you would like
audience • Have read all your earlier

papers
… • Thoroughly understand all the
relevant theory of cartesian
closed endomorphic bifunctors
• Are all eager to hear about the
latest developments in your
work
• Are fresh, alert, and ready for
action
Your The audience you get
actual • Have never heard of you

• Have heard of bi-functors, but
audience wish they hadn’t
… • Have just had lunch and are
ready for a doze
Your mission is to WAKE

THEM UP and make them glad
they did
What to put in
What to put in
1 2 3
Motivation Your key There is no
(20%) idea 3
(80%)
Motivatio They are thinking…
n • Why should I tune into this

talk?
• What is the problem?
You have two
minutes to engage • Why is it an interesting
your audience problem?
before they start to
doze. • Does this talk describe a
worthwhile advance?
Motivatio Example: Java class files are large (brief figures),
n and get sent over the network. Can we use
language-aware compression to shrink them? Yes,
and I’m going to show you how we can do 50%

You have 2 minutes
to answer these better than the best generic zipping technology
questions.
Don’t waste those
2 minutes. Example: Synchronisation errors in concurrent
programs are a nightmare to find. I’m going to
show you a type system that finds many such
errors at compile time.

Your key • You must identify a key idea.
”What I did this summer” is No Good.
idea • Be specific.
Don’t leave your audience to figure it out
If the audience for themselves.
remembers only • Be absolutely specific.
one thing from Say “If you remember nothing else,
your talk, what remember this.”
should it be?
• Organise your talk around this
specific goal.
Ruthlessly prune material that is irrelevant
to this goal.
Narrow, • Avoid shallow overviews at all
costs
deep • Cut to the chase: the technical
beats “meat”
• It’s ok to cover only part of your
wide, paper
shallow
N O!
YES!
Examples • To motivate the work
• To convey the basic intuition
are your • To illustrate The Idea in action
main • To show extreme cases
weapon • To highlight shortcomings
When time is short,

omit the general
case, not the
example.
Exception Solution 1: use data values to
carry exceptions
s in
Haskell?
Exceptions are to
Data Maybe a = nothing
| Just a
do with control
flow. There is no Lookup : : Name -> Dictionary -> Maybe Address
control flow in a
lazy functional
Often this is Just The Right Thing
program.
[Spivey 1990, Wadler “list of
successes”]
What to leave out
Outline of my talk
• Background
• The FLUGOL system
• Shortcomings of FLUGOL
• Overview of synthetic epimorphisms
• -reducible decidability of the pseudo-curried
fragment under the Snezkovwski invariant in
FLUGOL
• Benchmark results
• Related work
• Conclusions and further work
No “Outline of my talk”: conveys
near zero information at the start
outline! of your talk.
Worse, since your audience only
gives you 2 minutes before
dozing, you’ve just lost them
• But maybe put up an outline
for orientation after your
motivation
• …and signposts at pause points
during the talk
Related work
[PMW83] The seminal paper
[SPZ88]First use of epimorphisms
[PN93] Application of epimorphisms to
wibblification
[BXX98] Lacks full abstraction
[XXB99] Only runs on Sparc, no integration
with GUI
Do not But…
present • You absolutely must know the

related work; respond readily to
related questions
• Acknowledge co-authors (title slide),
work and pre-cursors (as you go along)
• Praise the opposition
“X’s very interesting work does Y; I
have extended it to do Z”
Technical detail
Omit • Even though every line is
drenched in your blood and
technical sweat, dense clouds of notation
details will send your audience to sleep
• Present specific aspects only;
refer to the paper for the details
• By all means have backup slides
to use in response to questions
Presenting your talk
How to present your talk
Your most potent weapon, by far,
is your
enthusiasm
!
Enthusias • If you do not seem excited by
your idea, why should the
m audience be?
• Enthusiasm makes people
dramatically more receptive
• It gets you loosened up,
breathing, moving around
Write (…or at least, polish it then)
• Your talk absolutely must be
your fresh in your mind
slides the • Ideas will occur to you during
night the conference, as you obsess
on your talk during other
before people’s presentations
Do not • “I didn’t have time to prepare
this talk properly”
apologise • “My computer broke down, so I
don’t have the results I
expected”
• “I don’t have time to tell you
about this”
• “I don’t feel qualified to address
this audience”
The jelly If you are anything like me, you
will experience apparently severe
effect pre-talk symptoms
• Inability to breathe
• Inability to stand up (legs give
way)
• Inability to operate brain
What to • Deep breathing during previous
talk
do about • Script your first few sentences
it precisely
You are not a wimp. (=> no brain required)
Everyone feels this • Move around a lot, use large
way. gestures, wave your arms,
stand on chairs
• Go to the loo first
Being • Face the audience, not the
screen
seen • Know your material
• Put your laptop in front of you,
screen towards you
• Don’t point much, but when you
do, point at the screen, not at
your laptop
Being • Speak to someone at the back
of the room, even if you have a
heard microphone on
• Make eye contact; identify a
nodder, and speak to him or her
(better still, more than one)
• Watch audience for questions…
Questions • Questions are not a problem.
Questions are a golden golden
golden opportunity to connect
with your audience.
• Specifically encourage
questions during your talk:
pause briefly now and then, ask
for questions
• Be prepared to truncate your
talk if you run out of time.
Better to connect, and not to
present all your material
Being a • Eye contact with speaker
• Nod frequently
good • Ask questions.
audience • Don’t wait for the speaker to invite
questions; ask
member • Start doing so the moment you lose
contact with the talk. The rest of the
audience will thank you for it.
• Stop when you sense that you are
beginning to de-rail the entire talk.
Presentin • Use a wireless presenter gizmo
• Test that your laptop works with
g your the projector, in advance
slides • Laptops break: leave a backup
copy on the web; bring a
backup copy on a disk or USB
key
Presentin A very annoying technique
g your • is to reveal
slides • your points
• one
• by one
• by one, unless…
• there is a punch line
Presentin Use animation effects
g your
slides very very
very very
very very
very
sparingly
Finishing • Audiences get restive and
essentially stop listening when
your time is up. Continuing is
very counter productive
Absolutely without • Simply truncate and conclude
fail, finish on time.
• Do not say “would you like me
to go on?” (it’s hard to say “no
thanks”)
Conclusion You will attend 50x as many talks
as you give. Watch other
: there is people’s talks intelligently, and
hope
The general
pick up ideas for what to do and
what to avoid.
standard is often
low.
You don’t have to
be outstanding to
stand out.
www.microsoft.com/research/people/simonpj
What your talk is for
Your paper Your talk Do not

= the beef = the beef confuse the
advertisemen two.
t
Do it! Do Good papers and talks are a
fundamental part of research
it! Do it! excellence
• Invest time
• Learn skills
Write a paper, and
give a talk, about • Practice
any idea, no matter
how weedy and
insignificant it may
seem to you.
Research Your papers and talks
is • Crystalise your ideas

• Communicate them to others
communicati
• Get feedback
on
The greatest ideas
are worthless if you • Build relationships
keep them to
yourself. • (And garner research brownie
points)
© Copyright Microsoft Corporation. All rights reserved.

How To Give A Great Research Talk

Uploaded by

Copyright:

Available Formats

How To Give A Great Research Talk

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

How To Give A Great Research Talk

Uploaded by

Copyright:

Available Formats

How to give a

Photo © James Millar/TEDxExeter

should • Think: how often have you said

audience • Have read all your earlier

actual • Have never heard of you

Your mission is to WAKE

n • Why should I tune into this

n and get sent over the network. Can we use

language-aware compression to shrink them? Yes,

and I’m going to show you how we can do 50%

programs are a nightmare to find. I’m going to

show you a type system that finds many such

errors at compile time.

When time is short,

present • You absolutely must know the

Your paper Your talk Do not

is • Crystalise your ideas

You might also like