This paper addresses the long-standing problem of inconsistently defined decentralization in computer systems by introducing a formal, graph-based ontology. This framework defines decentralization as a structural, relational, and subject-specific property, distinguishing it from distribution and enabling multi-dimensional evaluation with new metrics like Void Tolerance and Imperviousness.

This paper addresses the challenge of "hallucinations" in medical AI, particularly in surgical workflow recognition, by proposing a method to define and prevent topological errors. It demonstrates how integrating linear temporal logic predicates into probabilistic graphical models can mathematically guarantee the absence of certain critical errors, significantly improving accuracy and offering a pathway for robust AI regulation in medicine.

ControlRadio introduces a controllable generative framework that produces radio maps from natural language descriptions and environmental layouts. This approach demonstrates state-of-the-art accuracy and dramatically reduces computation time compared to conventional physical simulations, offering a new paradigm for wireless environment modeling.

This paper introduces ResidencyRL, a novel reinforcement learning framework designed to train clinical AI agents through extensive simulated multi-turn patient encounters. The method significantly improves diagnostic accuracy, management quality, and patient-centered communication, demonstrating robust and generalizable capabilities for sequential clinical decision-making.

This paper addresses the challenge of uniquely determining the causal order in probabilistic logic programs learned from data, where probabilistic information alone can lead to ambiguity. By exploiting the relationship with Bayesian networks and incorporating relational structure constraints, it provides conditions under which a unique causal order is established, thereby enabling well-defined intervention semantics.