Reports and presentation

Reports and presentations

Here are a few advices on our to prepare your final report and your presentation.

How to organize my report?

The general idea of a report is that a reader should be able to understand your result and easily continue your work without having to restart from scratch.

The usual way to organize a report is as follows:
  • The introduction shall recall the context and main objectives and problematics of your work

  • A background section may follow. It may give information about the company or lab in which you have made your work, and also the topic of your work.

  • A related work is then given, to explain what has already been tried by other companies or labs to solve your objectives. You should then reach the conclusion that nothing totally adapted to your problem has been proposed.

  • One or several sections must go deep into your contributions. Do explain why you have selected your solution. Only after you have explained the why, you can explain the "how": explain your algorithms, explain your code, and so on. Also, provide a way to reproduce your results. Last, be clear about what you have made on your own, and what was produced with colleagues or provided by the company or lab.

  • A discussion may follow to explain the ability or your contribution to address the objectives you have listed in introduction. This discussion may also be integrated into the conclusion.

  • The conclusion recalls which objectives have been reached, why some have not been reached and how they could be reached. Gives also possible extensions to your work.

  • I do not use a weight scale to evaluate your report. You should avoid useless blah-blah and focus on most important aspects in a consice way. Yet, don't forget the important information and the main objective of your report (being reused).

  • If you send me your report one day before your evaluation, or the day of your defense, I will not have read it. So you will have to wait for extra days to get your grade. Also, giving me your report late tends to drive me mad. So, be on time, do send it to me at least 3 work days before your defense.

Slides

A few advices for your slides
  • Do respect the timing.

  • Your presentation should not look like a movie of slides. Do spend at least 1 mn par slide, ideally 2.

  • You should not assume that the jury knows your objectives. So, do explain the context and the objectives of your work, before addressing contributions.

  • Do not put too much text on your slides: nobody can listen to you and read your slides at the same time. Do write only what is strictly necessary to support your speech.

  • A presentation does not need to contain all your contributions: it is usually better to focus on one of them rather than to present all of them in a superficial way. So, go deep into at least one contribution, e.g., show your code or system architecture, etc.

  • A demo is always appreciated. It makes your results more concrete. The demo can be provided as a movie if a live demo is not possible.

  • Basically, the outline of your presentation should follow the usual way of doing: context, problem and limitations of currently existing solutions, fast overview of your solution, zoom into at least one aspect of your solution, results, discussion and conclusion

  • Number your slides!, and put your name on the first slide

  • The smallest font should be 16. This applies to the text, figures, screen captures, code, .... If really some parts of your slides cannot be readable, either you have to zoom to these parts during the presentation, or you have to explain them (read the text, etc.).

Grading policy

For the report, the grading takes into account the conciseness, the reability, the respect of necessary sections, and the reusability of the report.

For the defense, I evaluate the quality of the slides, the respect of the timing, the understandability for non-expert of the domain, the respect of the necessary information (context, problematic, related work, contribution, discussion and conclusion).