Artificial Intelligence
Science as Refutation
Empirical Science
Theory, not Theorems
Empirical CS/AI
What are our hypotheses?
Why do experiments?
Basic issues in Experimental Design
Control
Case Study: MYCIN
Ceiling and Floor Effects
Ceiling Effects: Machine Learning
Ceiling Effects
Sampling Bias
Case Study: Phoenix
Sampling bias in Phoenix
Email: ipg@dcs.st-and.ac.uk
Home Page: http://www.dcs.st-and.ac.uk/~ipg
Download presentation source