Decentralized Online Convex Optimization (D-OCO) is a framework in which a set of local agents aims to minimize a sequence of global loss functions using only local computations and communications. The setting models distributed learning problems where agents must coordinate without a central server.
The new arXiv paper, "Decentralized Online Exp-Concave Optimization," applies this framework to exp-concave loss functions, as the title indicates. The abstract notes that recent work has addressed D-OCO, but the provided text is cut off mid-sentence, so the paper's specific contributions — regret bounds, algorithmic details, or communication guarantees — are not available in the source.
Because only one source was provided and its abstract is incomplete, this digest can report the problem setting but not the paper's results. No conflicting accounts exist to compare.