Researchers from the University of Edinburgh have proposed an ontological framework to define decentralization in computer science. They argue that current definitions are ambiguous and inconsistent, leading to difficulties in comparing different systems. The new framework, which uses a graph-based ontology, distinguishes decentralization from related concepts like distribution and provides two novel metrics for evaluation: Void Tolerance and Imperviousness. The authors claim their approach can be applied across various domains, including AI, blockchain, and IoT architectures. --- Why it matters: This work matters to researchers in AI because it addresses a long-standing problem in the field: defining decentralization. Current decentralized AI systems often rely on ambiguous definitions, which hinders formal analysis and comparison with other approaches. This framework provides a domain-independent foundation for evaluating decentralization, enabling more rigorous reasoning about communication architectures and protocol design. Source: https://arxiv.org/abs/2608.09748