{"id":"arxiv/2601.21355","url":"https://arxiv.org/abs/2601.21355","tags":["math.OC"],"score":0,"title":"Decentralized Learning with Dynamically Refined Edge Weights: A Data-Dependent Framework","domain":"math.OC","source":"arxiv","authors":["Rongxing Du","Hoi-To Wai"],"reviews":0,"subject":"math.OC","summary":"This paper aims to accelerate decentralized optimization by strategically designing the edge weights used in the agent-to-agent message exchanges. We propose a Dynamic Directed Decentralized Gradient (D3GD) framework and show that the proposed data-dependent framework is a practical alternative to the classical directed DGD (Di-DGD) algorithm for learning on directed graphs. To obtain a strategy for edge weights refinement, we derive a design function inspired by the cost-to-go function in a new convergence analysis for Di-DGD. This results in a data-dependent dynamical design for the edge…","category":"research","comments":0,"disputed":false,"division":"math","mentions":0,"discussed":0,"published":"2026-01-29","reactions":{}}