Lista de Disciplinas | CMP606
Tópicos Especiais em Computação DCVI – Scientic Methodology and Performance Evaluation for Experimental Computer Scientists
Professor responsável: Arnaud Legrand (CNRS, Grenoble, França) e Lucas Mello Schnorr
Carga Horária: 15 horas
1 créditos
Súmula
Reproducible research; Data presentation; Basic notion of statistics; Linear regression; Design of experiments
Objetivo
The aim of this course is to provide the fundamental basis for sound scientific methodology of performance evaluation of computer systems. Some of the topics include some maths (in particular probabilities and statistics) but I will adopt a very pragmatic presentation particularly suited to experimental computer scientists and will re-explain any notion that would not be clear. Every lecture will be backed up with practical sessions and worked out examples.
Conteúdo Programático
- introduce the audience to reproducible research and explain how to manage a laboratory notebook and report results (using tools like org-mode or knitR) . Such sound practice is necessary to allow other researchers (and even yourself) to build upon your work.
- provide an introduction to effective data presentation with R and illustrate with many classical pitfalls.
- provide the audience basic notions of statistics (esperance and variance estimation, confidence interval, distribution comparison, …) that are required in every-day life computer science experimentation.
- provide slightly more elaborate notions of statistics such as linear regression. The goal is that the audience understands such tool and the underlying assumptions to know when it makes sense or not.
- provide an introduction to the design of experiments, a fields of statistics that will help the audience to plan their experiments to obtain much more meaningful results with way less experiments.
A much more detailed version is available at the course website:
- https://github.com/alegrand/SMPE
Método de Trabalho
Cinco aulas expositivas de 100 minutos cada, totalizando 500 minutos. Os 400 minutos restantes para fechar 900 minutos (15 horas aula) serão realizados em formato educação à distância.
Bibliografia
- Measuring Computer Performance: A Practitioner’s Guide by David Lijla. Cambridge University Press, 2005
- The Art of Computer Systems Performance Analysis by Raj Jain. John Wiley & Sons 1991.