Research method in AI: Reproducibility of results
Master thesis
Permanent lenke
http://hdl.handle.net/11250/2571694Utgivelsesdato
2018Metadata
Vis full innførselSamlinger
Sammendrag
In recent years, increasing attention has been given to anapparent issue within the scientific fields where the reproducibility ofscientific results is questioned. This trend is commonly termed a crisis of reproducibility, or a reproducibility crisis. While medicineis a central scientific field here, it spans most other fields as well,including the computational sciences.
This thesis investigates the reproducibility of recent and centralscientific publications within artifical intelligence, through aselection of the most cited papers the past few years. The reproducibilityis investigated by attempting to reproduce the presented results, basedon the description of the method presented in the paper.