News
The Ultimate Guide to Forgetting
Automated reasoning in Artificial Intelligence builds on techniques to leverage acquired knowledge in useful ways. However, as also happens with humans, a good reasoner must selectively forget information that only hinders effectiveness, while preserving what matters the most! How to do that in the right way poses significant challenges to the research in the field.
Many different approaches have been proposed to remove or hide information from a knowledge-base that is no longer considered relevant, resulting in a complex and confusing set of alternatives to perform this operation.
Recently, in their KR 2016 paper “The Ultimate Guide to Forgetting in Answer-Set Programming”, Ricardo Gonçalves, Matthias Knorr and João Leite, NOVA LINCS researchers from the Knowledge-Based Systems Group, thoroughly examined existing approaches for forgetting in rule-based knowledge bases, drawing a complete picture of the landscape, which includes many novel, even surprising, results, thus providing a guide to help users in choosing the most adequate approach for their application requirements.
This paper will be presented at 15th International Conference on Principles of Knowledge Representation and Reasoning (KR 2016) on April 25-29, 2016. KR is the world leading conference on Knowledge Representation and Reasoning.