SCM Literature

Course literature consists of the following two books, handouts given in class, and web pages.

[Cohen] Paul Cohen: Empirical Methods for Artificial Intelligence. MIT Press, ISBN 0-262-03225-2, 1995 (mandatory)
[Dawson] Christian W. Dawson: Projects in Computing and Information Systems. Addison Wesley, ISBN 978-0-273-72131-4, 2009, 2nd edition (not mandatory)

The two books are available at the university bookshop. The first (Cohen) is marked mandatory because you will have a very hard time following the course if you do not have it. The second (Dawson) is not mandatory because you can understand most of the course without it; it will however most likely be very useful for you in the future and will also make it easier to follow the course.

A common request was to get more details on the statistics, for this reason I am collecting alternate resources on the topic that can provide a quick overview/introduction. If you find something relevant, please send Ulrik an email. Currently available material:

Statistics cheat sheet: a brief overview of relevant statistics formulae
Wikipedia: overall description plus list of relevant statistics formulae
Mathematica: download Mathematica player to try out some of these distributions interactively. Note that the full Mathematica is installed on machines in the terminal room.

Additional literature

SCM-01 An insightful comparison between experiments in mobile robotics and in science by F Amigoni, M Reggiani, V Schiaffonati - Autonomous Robots 2009 - Springer [local copy], [external link]
SCM-02 Ownership types for object encapsulation by Chandrasekhar Boyapati, Barbara Liskov, Liuba Shrira - POPL'03 - ACM [local copy] [external link]
SCM-03 Ownership types for flexible alias protection by David G. Clarke, John M. Potter, James Noble - OOPSLA'98 - ACM [local copy] [external link]
SCM-04 On the Impact of Aspectual Decompositions on Design Stability: An Empirical Study by Phil Greenwood, Thiago Bartolomei, Eduardo Figueiredo, Marcos Dosea, Alessandro Garcia, Nelio Cacho, Claudio Sant'Anna, Sergio Soares, Paulo Borba, Uira Kulesza, and Awais Rashid - ECOOP'07 - Springer [local copy] [external link]

Survey paper examples

The survey project concerns reading, summarizing, and analyzing a number of papers from the point of view of empirical methods. Survey papers are often written with a particular point of view in mind; SCM-01 above contains a survey written with the point of view of assessing the use of empirical methods in robotics. Below is a number of other survey papers that can serve as inspiration; keep in mind the notes that I have added next to each paper, they describe the overall style of the survey paper.

Survey on domain-specific languages: the goal of this survey paper is to provide a structured view of domain-specific languages and moreover to provide an insight into when it is relevant to develop domain-specific languages and how it is normally done. There is in this paper a strong "decomposition" into areas specific to the domain; each reference may be used in different parts of the survey. Your survey could be structured similarly depending on the topic you have chosen; it would be divided into a number of sections that address specific subtopics relevant to all papers, ending with an assessment of the use of empirical methods.
Survey on facial recognition: this survey paper provides what appears to be a very comprehensive overview of the literature on facial recognition and gives a detailed description of the methods found in selected key papers, and moreover provides some more general analysis at the end. Your survey could be structured similarly: Section 2 provides some general background, whereas Sections 3 and 4 review specific papers, and the subsequent sections provide in-depth analysis of specific topics (your topic would be the use of empirical methods).

Numerous surveys can be found online, a favorite of Ulrik is the ACM Computing Surveys (access requires going to the ACM digitial library through the SDU library link).

Web resources

An important part of this course is learning how to read research papers and present them in a useful way, both in written form and orally. The following links contain generally useful material on this issue:

Research and scientific hypotheses: [link], [link]
How to read scientific papers: [link], [link], (more)
How to present scientific papers: [link:pdf]
Report writing: [link], [link], [link]

Finding research papers is much easier if you know where to look. Many resources are available from the SDU library web page; here are a number of Ulrik's favorite search engines:

ACM Digital Library: Primarily indexes ACM publications (good for programming languages and more theoretical issues), Ulrik often gets good results here with many cross-references.
IEEE Explore: Primarily indexes IEEE publications (good for software engineering and more engineering-oriented issues), Ulrik often has a harder time getting good results, but depending on the domain this is the place to get certain papers.
Citeseer: Indexes all computer science publications, very comprehensive. Ulrik sometimes is disappointed with the search engine when doing "wide" searches. Useable outside university network, often contains cached copies of papers.
Google Scholar: Indexes all computer science publications. Ulrik tried it a while ago and was not so impressed, but it may have improved in the meantime.

Note that some of these search engines (ACM and IEEE) only can be accessed from within SDU's network.

I encourage you to use LaTeX and BibTeX for your report, see latex-project.org. Click "Getting LaTeX" and then select your distribution. There are numerous tutorials available, e.g. the LaTeX primer. A tiny example of a LaTeX document is available under materials/latex.

Curriculum

The exam is a written report that will have Cohen and Dawson (above) as curriculum, but the details of the curriculum will be finalized 15th of October.