Modularity and Splitting Techniques for Knowledge Representation and Reasoning

Call for Papers

We consider modularity and splitting techniques to be a cross-cutting issue being basically relevant for all subareas of KR, in the tradition of divide-and-conquer methodologies. In particular, researchers from nonmonotonic reasoning and belief revision and from answer set programming might be interested in this topic, but we also welcome submissions from more classical subfields like description logics, or using quantitative (e.g., probabilistics) and network-based methodologies.

Topics of the workshop include, but are not limited to

• Modular representations of knowledge and belief bases
• Syntax and semantic splittings
• Global vs. local knowledge representation and reasoning
• Decomposition and composition
• Merging of knowledge and beliefs
• Graph- and network-based representations
• Relevance and independence
• Forgetting and abstraction
• KR and cognition

See "Submission" for instructions on how to submit your work.

About the Workshop

The field of knowledge representation and reasoning (KR) has brought forth a great variety of sophisticated approaches to represent knowledge and beliefs of humans and agents, and to reason from such representations. While KR formalisms usually are explicitly designed to model real-life scenarios including commonsense knowledge, the complexity of such situations often poses a severe challenge to reasoning efficiently in these formalisms.

A key to break up the complexity of information is to represent and reason in local contexts. Ideally, these local contexts should contain all information relevant to a specific aspect of the problem under consideration, abstracting from irrelevant details, so that the outcomes of local reasoning procedures after merging coincide with what would have been obtained theoretically in the global picture. Examples of such techniques have been proposed for nonmonotonic reasoning and belief revision under the name syntax splitting, and in modular answer set programming where modules take the role of local contexts. Networks can provide structures to compose the global picture from local contexts, like in probabilistic networks. Interestingly, humans are performing quite well in their everyday lives by recognizing relevant information quickly and reason from local contexts successfully with their limited resources. So, also cognitive aspects of knowledge representation and reasoning might be relevant.

The aim of the workshop is to bring together researchers from different (sub)fields in KR and related to KR to discuss modularity and splitting techniques for knowledge representation and reasoning. We invite papers addressing these issues from all aspects of KR, being based on different methodologies using symbolic, qualitative, or quantitative approaches. Discussing modularity and splitting in a broad scientific context shall help understand its underlying mechanisms better and foster collaborations across scientific domains.