Towards Finding Animal Replacement Methods

Abstract

Protecting animal rights and reducing animal suffering in experimentation is a globally recognized goal in science. Yet numbers have been rising, especially in basic research. While most scientists agree that they would prefer to use less invasive methods, studies have shown that current information systems are not equipped to support the search for alternative methods. In this paper, we outline our investigations into the problem. We look into supervised and semi-supervised methods and outline ways to remedy the problem. We learned that machine assisted methods can identify the documents in question, but they are not perfect yet and in particular the question about gathering sufficient training data is unsolved.