how_to_give_talks.qmd

# How to Give Talks {#sec-how_to_give_talks}

## Introduction

Giving good talks is important for people who use the material of this
book--for students, researchers, and engineers. One might think, "It is
the work itself that really counts. Giving a talk about it is
secondary." But the ability to give a good talk is like having a big
serve in tennis---by itself, it doesn't win the game for you, but it
sure helps. And the very best tennis players all have great serves. So
we include in this chapter material that we have found helpful in giving
talks.

There are many other sources for giving talks, and you should access
several of them to pick out the tips that work best for you. Prof.
Patrick Winston's talk about giving talks is great @WinstonSpeaks, and
another good source is this book @Hoogterp. This chapter has heuristics
that have worked well for us.

### A Taxonomy of Talks

There are different kinds of talks. Conference and workshop talks may
range in length between 3 minutes for a "fast-forward" talk to 45 or 60
minutes for a keynote address. You may be one of many speakers,
sometimes speaking in different rooms at the same time. Sometimes you
will visit a university or company and speak in a special seminar. (Here
@MITFacultyJobTalks2022 is a list of tips regarding faculty job talks
from the MIT faculty.)

It is very often the case that the audience will have a wide range of
familiarity with the work and topic that you will present about. In a
small seminar to a competing research group, there may be many people
who are very familiar with your work. But even in such a small seminar,
and especially to a larger group, there may be many people who don't
know your area, nor perhaps not even your broad area. So you will to
include material to let the uninitiated not be lost. Even people who are
very familiar with the topic will appreciate hearing material they
already understand presented clearly and well.

The tips in the chapter apply to talks of all lengths. However, one
class of talks is so short as to be a special case: the very short talk.

## Very Short Talks (2 -- 10 minutes)

Short talks are often advertisements, designed to persuade the listener
to read your paper, or to listen to a longer version of the talk. Rather
than trying to convey many details about your work, you should aim to
have the audience remember the main idea behind your work, and to share
your excitement about the work.

For five minute talks describing students' final project class
presentations, we suggest that the students cover these points:

1.  What problem did you address?

2.  Why is it interesting?

3.  Why is it hard?

4.  What was the key to your approach?

5.  How well did it work?

That structure can be used for other short talks, too. For very short
talks, the time needed to go through the whole thing is minimal, and
there is no excuse not to practise the talk, from start to finish,
several times. Next, we cover points that we feel relate to all types of
technical talks, regardless of length.

## Preparation



![The key to a good talk:  preparation and practice.](figures/how_to_give_talks/speaker.jpg){width="80%" #fig-speaker}

We believe the keys to both a good talk and to overcoming nerves are the
same: prepare and practice. Think through your talk. It's almost always
better to give a talk from notes than to read it from a script--the
audience feels you're speaking to them so it's easier for them to pay
attention. But for people giving a talk in a language that they are not
comfortable with, writing out a script ahead of time may be a better
solution. There may be parts of the talk that you feel are difficult to
get through. For those parts, even for a native speaker, it may help to
write out what you plan to say. When you give the talk, you may ignore
the script, but having written it out once will make it easier to say.

Think through the talk and find the story--how one part relates to the
next. If you can't find that story, you may want to reorder the
presentation to create a good story.

When you can, you should visit the room you'll be speaking in to
identify any issues that may come up. Will you be blocking anyone as you
stand at the lectern to give the talk? Will there be someone there to
help you set up? You should decide how to position yourself so that you
can see your presentation, and also engage well with the audience.

## Nervousness

One of us writes, "I used to be a nervous public speaker. I still find
the hours before giving a presentation to be stressful. I'm always
reviewing my talk to try to improve it or to track down the answers to
questions I think someone might ask me. The more I prepare, and the
better I know the lecture material, the less stressed I am.\"

A great cure for nervousness is to practice the talk aloud. Practice by
yourself. Practice in front of your friends.

One approach @Hoogterp to calming nervousness is to remind yourself,
"Get over it. They're not there to see you, they're there to hear the
information. Just convey the information to them."

Some find this quote, attributed to Dr. Rob Gilbert @Gilbert, to be
helpful: "It's all right to have butterflies in your stomach. Your job
is to get them to fly in formation."

## Your Distracted Audience

Before giving a talk, your mental image of the audience may be of many
people, listening to your every word. In reality, on most every speaking
situation, your audience is a collection of people who are checking
their social media or email, looking on their laptops for other talks to
attend, or are hungry or fidgety. How does one speak to such an
audience? We have to practise and prepare in order to engage the
audience.

## Ways to Engage the Audience

In the middle of Patrick Winston's talk about public speaking
@WinstonSpeaks, he asks the audience the question, "Can you think of a
technique to get the audience more engaged in the talk?" The answer, of
course, is to ask them questions.

Bill Hoogterp @Hoogterp expands on that: "The audience is like a sheep
dog, always wanting to be working. Give the audience something to work
on while they're listening to you. 'Four' pushes people back, while 'two
plus two' draws them in.\" You might pose a sub-problem to them that you
had to address in your work, to get them thinking about how to solve
that problem. Then they'll be more receptive to hear your solution.

### Layer Your Talk

It's often easier for the audience to listen to your talk if you layer
it---provide verbal cues for new concepts and transitions between
sections of the talk. This helps even the attentive listener to follow
the structure of your talk. It keeps the distracted listener from being
completely lost. The distracted listener, only listening for the verbal
headings of your talk, may hear, "The probability of an observation has
three terms to it. $\ldots$ blah blah blah blah $\ldots$ So that gives
us the objective function we want to optimize. Now, how do we find the
optimal value? There are two approaches you can take. $\ldots$ blah blah
blah blah So now, with these tools in hand, we can apply this methods to
real images. $\ldots$ blah blah blah blah $\ldots$\" Your verbal cues
can help even a distracted listener follow your main points.

You can also make your talk easier to follow if you add verbal
dynamics---variations in speed or intensity. Find a part of the talk
that is particularly special to you and let that show through. One of us
writes, \"When I gave talks about image deblurring, I would emphasize
one aspect that I particularly enjoy: 'I love this problem; it's
beautifully underdetermined. There are many ways we can explain a blurry
image. It could be that that's what was there in the world--- we took a
sharp picture of a world that happened to look blurry. Or we took a
blurred image of a sharp world.' \" The audience loves to watch you be
excited about something!

### People Like to See a Good Fight

As described by @Adelson95, the audience loves to watch a good fight. You can set up a fight between two competing conjectures. For example, you might say, "The flat earth theory predicts that ships will appear on the horizon as small versions of the complete ship. Under that theory, you’d expect approaching ships to appear as in @fig-boats-a. Conversely, the round earth theory predicts that the top of the sails will appear first, then gradually the rest of the ship below it, resulting in approaching ships looking as in @fig-boats-b. The audience waits with anticipation as you show them the result from your experiment, @fig-boats-c, thus revealing the winning theory."

::: {#fig-boats layout-ncol=3}

![The appearance of ships approaching a port, from the flat-earth theory.](figures/how_to_give_talks/boatsflat.jpg){#fig-boats-a width=100%}

![Their appearance as predicted by the round-earth theory.](figures/how_to_give_talks/boatsround.jpg){#fig-boats-b width=100%}

![The experimental evidence: a photograph of ships approaching a port.](figures/how_to_give_talks/boatsphoto.jpg){#fig-boats-c width=100%}

Fig (a) The appearance of ships approaching a port, from the flat-earth theory. (b) Their appearance as predicted by the round-earth theory. (c) The experimental evidence: a photograph of ships approaching a port @mnsomero
:::

## Show Yourself to the Audience

In a lovely video, the actor Alan Alda describes the importance of
connecting with the audience in scientific presentations @AlanAlda. Alda
noted the improvement in scientific speaking of volunteers after they
engaged in improvisational theater exercises. After the exercises, the
volunteers were primed to engage and connect with others, and their
scientific talks sparkled.

Here's what we think an audience wants from a technical talk: (a) To
have one part follow from another and make sense. (b) To learn a few
things. (c) To connect with the speaker, to share their excitement for
the topic. They want to watch you love something!

### How to End the Talk

How should you end your talk? A common, but awkward, way is to end by
asking, "Are there any questions?" The audience wants to clap at the end
of a talk, but with that ending, they don't know whether to applaud or
to ask a question, and they're likely to give only scattered applause
@Adelson95. Better is to close with "Thank you." The audience then knows
it's their time to applaud, and they will. Then you can ask for
questions.

We note that Patrick Winston disagrees with ending a talk with \"thank
you\" @WinstonSpeaks. He didn't like the implication that the audience
was doing you some favor for attending your talk, and preferred to
remind the audience of how he has kept his earlier promise to them
regarding what they would learn after they listened to his talk.

## Concluding Remarks

Prepare and practice your talks! As you give the talk, let the audience
see how much you enjoy what you've worked on. Follow-up with the
references in this chapter to find the collection of talk tips that
works best for you. Preparation and rehearsal are the best cures we know
of for nervousness about giving a talk.