|
Instructors: Dr. Sven Magg
Event type:
Lecture
Displayed in timetable as:
RM - VL
Hours per week:
2
Credits:
3,0
Language of instruction:
English
Min. | Max. participants:
- | 25
Comments/contents:
With the growing complexity of computer programs and intelligent systems, quantitative methods to analyse data are becoming increasingly important. Students of intelligent, adaptive systems are nowadays expected to be able to formulate hypotheses on the behaviour of artificial systems, create well-defined experiments, gather empirical data, and appropriately analyse those data to draw conclusions.
In this lecture we will introduce students to the core concepts of the scientific process, beginning at experiment design and execution all the way to data analysis and publication. A focus will be on methods and tools useful for the field of computer science and artificial intelligence. Compared to other natural sciences, the test subjects in computer science are often artificial systems or humans that interact with them and the lecture provides an overview over techniques frequently used especially in this domain.
Topics will include types and design of empirical studies and their application areas as well as statistical methods for the analysis of different forms of qualitative and quantitative data, including methods like Monte-Carlo Sampling and Bootstrapping.
Learning objectives:
The objective of this lecture is to gain a deeper understanding of the scientific methods and their application in the field of computer science and artificial intelligence:
• Learning the basic principles of scientific processes
• Gain applicable knowledge on experiment design and execution
• Understand hypothesis testing and statistical methods
Didactic concept:
The interactive lecture will be tightly coupled with a practical seminar and will equip students with the theoretical background and concepts. Those will then be deepened in group discussions, practical experiments and data analysis. Over the course of the lecture and the exercises, students are going to design and execute at least one experiment and analyse the data followed by a presentation and discussion of their approach and results. Practical implementation of methods for calculation and visualization in Python will help to deepen the knowledge.
The lecture 20/21 will be held online over video chat and collaboration tools due to changes because of the Corona virus. The exact procedure and implementation will be discussed with the students in the first lecture.
Literature:
• Cohen, P. R. Empirical methods for artificial intelligence. MIT Press, Cambridge, Mass. 1995.
• Field, A., Miles, J., Field, Z. Discovering statistics using R. SAGE, Los Angeles, 2012.
• Allen B. Downey., Think Stats 2e. Green Tea Press Books, 2014. (Freely available online)
More literature will be recommended during the lecture.
|
