CS888 - Advanced Topics in Computer Graphics

Making a Good Presentation

Presenting a paper in a grad class seems pretty simple: you read the paper; you start making some slides; you read the paper again; you finish making your slides; and then you go to class and talk about what you put on your slides. Yet, even one term of sitting in the audience tells you emphatically that most presentations are not very good. Smart, well-educated people are doing something simple badly: what gives? The hard parts are easy to identify: reading the paper, making the slides, and talking about what's in the slides.

Reading the Paper

Ask yourself: what is the central idea of this paper? This can be hard: often the author doesn't know. But it's the central idea that gives the paper a relevance that goes beyond the immediate circumstances of its creation, all too often not much more than the advancement of its authors' careers.

Many students think that they show how smart they are by discovering things that are wrong in the paper they chose. For me, this is not news: I have read and reviewed hundreds of papers and almost every one of them had something wrong*. Things that are wrong are reasons to disregard the paper, ipso facto, to pay no attention to the presentation. What will affect my, and others', lives is what's right about the paper. Something made the author(s) think the paper worth all the time and hassle to write. What was it? It should be a deep idea, possibly one the expression of which is spread over several papers. The purpose of reading is not to nitpick faults but to find the great ideas, which are often hidden.

Great ideas are almost always simple. The more simple an idea is the more wide its applicability. Keep reading and thinking until the ideas seem simple to you: when they start seeming obvious you know you are on the right track. In the best presentations the ideas presented are unfamiliar to the audience at the beginning, and obvious at the end.

Making the Slides

There is a rule of thumb that it's impossible for an audience to absorb more than one truly new idea every fifteen minutes. If you believe this -- I do! -- the the most important thing to do in creating a presentation is to identify the two points you wish to make before you try to do anything else. Write them down, then reread the paper, and other papers derived from it, filling in all the details about them. Such as,

You can structure your presentation in the following way.

  1. Introduction, in which you describe the two ideas, giving anecdotes.
  2. Development, in which you describe the history, logic, evidence, consequences, and so on of the ideas. It's best if you don't do this serially, first one idea then the other. Given that they are in the same paper/presentation they should have a strong interrealtionship. This is your opportunity to bring this out.
  3. Conclusion, in which you review the ideas in a deeper way than you introduced them.

Those of you who know western music will notice the sonata form, which is the basis of classical (1760-1830) and post classical music. Obviously, a cadenza or coda drawing on your personal response to the ideas is an excellent addition.

On the role of pictures

This is a course on visualization. Show the audience you are an expert on visualization by using pictures and diagrams that illustrate what you are explaining.

My favourite presentations are the ones in which almost everything on almost all the slides is pictorial, with the presenter talking my through the pictures as the dominant method of explaining his or her ideas.

Please, keep cute clip art to a minimum**.

Giving the Presentation


Interesting presentations work better than uninteresting ones: the audience actually listens, which certainly increases the chance they will think and remember.

Also, while making jokes is one way of being interesting, it's not the only one.

* I do not read or review papers I write. :-)

** If you are unsure whether or not a piece of clip art is cute, it probably is.

Return to: