Melvin's digital garden

Presentation Tips

Thesis presentation tips from HW

Different from a conference presentation as evaluators are trying to grade your work.

Motivation, related work - Show general understanding. What is my niche (thesis) ?

Show some details - Where is the meat? Select section with some depth/novelty

Data, validation - Details of the data, methodology and results

Tips from Steve Job’s MacWorld keynote

Set the theme. “There is something in the air today.”

Demonstrate enthusiasm. “extraordinary,” “amazing,” and “cool”

Provide an outline. “There are four things I want to talk about today. So let’s get started…”

Make numbers meaningful.

Try for an unforgettable moment.

Create visual slides.

Give ‘em a show.

Don’t sweat the small stuff.

Sell the benefit.

Rehearse, rehearse, rehearse.

Flow diagram

use a flow diagram to given an overview of the procedure

make use of different colors to indicate which components are from related work and which are novel


convex hull as a rubber band


make a connection to earlier talks, to the audience

at most 1/4 of talk on related work

have something easy and something substantial

don’t go back and forth between slides, duplicate slides if necessary

have practice runs, read aloud what to say for each slide

break into several sections, pause to check if on time

emphasize the central/main idea, in simple English

use diagrams/pictures to show key ideas

every slides show show only ONE story, be focused

show current results even if it did not make it into paper

notational consistency

for each slide, know what is the key message

take home message should be 2~3 key points

highlight the potential/good points of the method, in introduction, in solution, give strong evidence in experimental results


give contextual cues in poster to guide the reader

use “slides” in the poster to group together related material


explain how data is obtained

show examples where your method does better

give comments on interesting parts, highlight important areas


put where result is quoted


choose suitable examples to illustrate the point

use a running example


top down preferred over bottom up

what is the objective?

why do we need this?

general idea

how is it done?

details if necessary


explain the basics (concepts, terminology)

anticipate the questions

relevance/benefits/target users

questions to be solved

what is the area, issues, challenges


steps to take

literature review

note down all the points that are to be presented

know the purpose of each slide, be focus, one slide one main point

tone of presentation

identify the difficult/confusing parts

repeat the data if necessary for comparison

How to present a paper by Leslie Lamport, 4 August 1979

\subsection{WHAT TO SAY} Don’t give your paper; the audience can’t take it. If someone can understand in thirty minutes what it took you weeks to develop, then you’re in the wrong business.

Do advertise your paper. The purpose of an automobile ad is to get potential customers to the showroom, not to give technical specifications. The purpose of your talk is to get people who might be interested in your work to read the paper, not to save them the trouble of reading it.

Giving a good presentation is an art, requiring both practice and talent. No rules can turn you into an artist, but the following suggestions might be helpful.

Describe simple examples rather than general results. Try to make the examples much too simple – you will not succeed.

Don’t use formalism. If your results cannot be described simply and informally, then there is no reason why anyone should be interested in them.

It is better to be inaccurate than incomprehensible. The place for accuracy is in the paper. (However, false advertising is unethical.)

\subsection{HOW TO SAY IT} Slides are effective. Here are some suggestions for their proper use.

Don’t put too much on a slide – a picture of a thousand words is worthless. For 8 x 11 slides, all letters should be at least 3/8 inch high, with plenty of blank space. People in the back row have to read them too.

Slides should be neat and legible. The listener isn’t your secretary; it’s not his job to decipher your handwriting.

A rapid sequence of slides has a hypnotic effect. Unless you are a licensed hypnotist, don’t use more than one slide per minute.

Time your talk. Running over your allotted time is a mark of incompetence, and displaying your incompetence is a poor way to get someone to read your paper. Remember that talking to an audience takes longer than talking to a mirror.

Links to this note