[Reading Notes] Slide Design for Research Presentations

Slide Design for Research Presentations https://www.kodansha.co.jp/book/products/0000194774

Conclusion, reason, transition, conclusion

In practice, in the examples, it feels more like conclusion → reason/supplement → conclusion.

Example

Question: Why have you not written a single paper this year?

Answer: I do not have time to write papers.

This is because writing a paper involves many steps. The final quality changes depending on how much time can be invested in each one of those steps. Therefore, if you want to write a good paper, you must invest an appropriate amount of time.

The reason I have not written a single paper this year is that I have not yet been able to secure that much time.

Write it in six lines. Six paragraphs?

If you change the premise of a slide, there may be content you can cut.
- Premise: limiting the scope, for example.

Example

BAD

Load-balancing branching plan

What to do if a problem occurs → Do nothing
When a problem occurs → Do XX

GOOD

Review of the load-balancing branching plan when problems occur

Problem 1 → Do XX
Problem 2 → Do YY

Where to use animation

Guiding the audience's gaze
- Limits the information they see
Developing a story within a single slide (like a picture-story show)
- Makes it clear what is being discussed “now”?
Example: for a graph, gradually increase the types of data (with bar charts, line graphs, etc.)
Use animation to explain mechanisms.
- LLM tokenization, vectorization, and so on.
Formulas
Having animation on about one out of every three or four slides is good.
- Add “variation in pace” to the presentation.

Make slide titles descriptive (titles that communicate)

Assuming that a graph is being displayed,
- Results → X is becoming ~
- Comparison of X and Y → X is greater in the case of ~
Is it best to have each section on one slide and use it when transitioning?

About sizes

At least 20 pt; ideally 30 pt or more; titles should be 36 pt or more.
- In the example, titles are 44 pt and body text is 36 pt.

Do not quote as-is.
- For example, when photos are arranged so that a trend can be understood over time.
  - If there are many, each one becomes smaller.
  - The audience also has to follow them with their eyes.
  → Use the first and last, or the first, middle, and last, for example.

About terminology

It is not enough to say that you explained it at the beginning. No one remembers it.
- For example, other than terminology, experimental groups.
  - At the end of the experimental section, when saying, “In other words, we found that Group A is effective for X,” if you only show “Group A,” the audience will think, “What was Group A again?”

→ Basically, it is better not to use abbreviations.

Proceed by replacing terms with Japanese expressions.
Turn definitions into diagrams.
- The order of LLM processing is tokenization, vectorization, inference by a neural network?, contextual understanding, and decoding.
  - Make this series of steps into a diagram, for example. Not just a text-only flowchart.
When the formal name becomes long,
- Explain the abbreviation verbally.
Do not put on airs.
The balance between what needs explanation and what does not.
- Know, Know, Know, Don’t know

Figures

Display formulas together with figures, for example.
- Explaining terms, etc.
  - Example: lean body mass. Human body - body fat mass = lean body mass. (Why Your Presentation Does Not Communicate, p. 88)
Remove items unnecessary for the message.
- Descriptive title → what you want to communicate
  - A figure that makes what you want to communicate convincing.

Talk

State what you want to communicate, restate the title in simpler terms, and show where you are going next.
- At the beginning of each section, say what you are about to discuss and where you are heading.

Example

“Safety and efficacy of X”

In this presentation, I hope you will learn that X, a new method for Y, is both safe and effective.

Images of the case

In this case, we performed X and had the experience of Y. Here, let us look at how Y came about.

The introduction should be about one-third to one-quarter of the whole.
Improve sentences based on how they sound when spoken.
- Overly roundabout, difficult to say, and so on.
- Add line breaks or change the wording.
Make the audience feel that it is their problem, that it relates to them; involve them.
- Have them raise their hands or take action.
- Include “everyone,” or a name for the audience, in what you say.

Example

Some of you may be doctors, so I am sure you have had a cold sweat during surgery when you could not find the lesion that needed to be removed. X, which I will discuss today, solves this very troubling problem for us.

As you all know, these days X often leads to Y, right? So we do A, but the actual problem is Z.

You may wonder, “Why, when it is R?” But this is S.

When it becomes Y, we engineers are in trouble.

What solves that problem is XXX, which I will talk about today.

When the audience is the general public

Some of you may know someone around you who has developed X. X is currently the most common ~.

Ideally, it would be detected and treated early, and recently CT scans in comprehensive medical checkups can detect lesions with diameters of ◯ centimeters or less.

However, there is a major problem here. Even if early-stage tumors are found through comprehensive medical checkups, it can become impossible to tell where the tumor is during surgery.

In severe cases, they say, “Let’s let it grow a little more before removing it.” That defeats the purpose of early detection, does it not?

That is where X, which I will discuss today, comes in. With this method, lesions can be removed almost certainly, and without removing too much or too little.

People who have and have not given an English presentation (English presentation)

Have

I think all of you have some experience giving research presentations in English. How was it? Was it easy? For many of you here, preparation was probably difficult, and the presentation itself likely felt different from presenting in Japanese. Above all, the Q&A may have been downright frightening.

There are many reasons for this. And this is good news: there are several measures you can take right away. Today, we will look at measures with immediate effects that will help even if you have an English presentation tomorrow, as well as measures you can take from a long-term perspective.

Have not

Do you all think that giving a presentation in English is something far off in the future? That is understandable: you need a certain amount of experience before you can have a topic suitable for presenting at an international conference.

However, what I will discuss today is not at all too early for you. Rather, it is something that will be useful in the near future if you hear it now. Conversely, your senior colleagues who did not have an opportunity to hear this kind of talk have generally struggled greatly with English presentations. There are measures you can take even once an English presentation is approaching, but they have limits. If you aim to be active internationally as a researcher, you should start honing your English presentation skills now.

Agenda

If the presentation genre is recognized, the flow can be inferred, so write the content of the sections.

For example, for a research presentation, everyone knows it will be background → methods → results → conclusion, so do not make those the agenda items.

→ That said, there may be situations where writing them makes things easier to understand, so perhaps listing both is an option...? (Consider)

Example

Why can’t you write a paper?

Why can’t you write? Consider measures based on the causes.
Timing and tips for writing papers in the cancer field
Writing alone is not enough: how to complete the long journey to publication

How to present data

Even with the same data, the presentation method changes things. Prepare the data.
- In comparisons, the calculation formula changes how it is shown. (What to use as the denominator, which side to use as the denominator.)

Example

Data to show the declining birthrate and aging population

y: population, x: time series

In recent years, if both young people and older people are increasing, it is not persuasive.

“Is it really a declining birthrate and aging population? Young people are increasing too.”

In that case, use a ratio.

Showing the ratio of the older population to the younger population makes the graph's changes more persuasive.

Outlook (Limitation: limitations)

In Japanese, perhaps “issues”...?

It may be good to display solutions corresponding to limitations (problems) with animation.
If there is a plan for improvement, present milestones. (p. 109)

Conclusion (final slide)

About two sentences. Content that takes about 20 seconds to say.
- Two short sentences rather than one long sentence.
Do not display it only briefly. It will not remain in the audience's minds.
- You should spend more time on it than on other parts.
It may be good to add a Summary slide before the Conclusion.
- Even better if it is summarized in a conceptual diagram!

Acknowledgments

Not needed. It is better to end with the Conclusion so that it stays in the audience's minds.

Besides, the audience is not interested.

If you want to acknowledge contributors, show them when presenting the materials they contributed to. (Photo, text?)

If you do not have time, narrow it down rather than speaking quickly

There are all these things related to this, but due to time constraints, today I will talk only about these particularly important ones.
- Example: “There are three techniques like this, but due to time constraints, today I will talk only about the particularly important X.”

How to speak

Notice your habits, then stop them. Do not fear pauses.

Fillers, verbal tics, and so on.
- They are annoying to listeners, so stop.
Pauses are actually good. They give the audience time to understand.

Lower your pitch and speak from your diaphragm.

A confident way of speaking

Do not speak too smoothly.

Add variation in pace (silence).
- That is what drinking water is for!