The paper investigates how scaling the number of clients in decentralized learning affects information propagation, consensus, and accuracy. By comparing a degree‑two Ring, a Static Random graph, and a locally adaptive Local‑First Heuristic Evolution (LFHE) topology on CIFAR‑10, the authors show that the Ring’s spectral gap decays as Θ(N⁻²), leading to slower convergence and higher disagreement. Experiments reveal that keeping local dataset size constant mitigates population penalties, and that LFHE can improve accuracy and consensus at the cost of increased communication.
arXiv:2610.08176v1 Announce Type: new Abstract: Decentralized learning is highly sensitive to communication topology under non-IID data. Adaptive peer-selection methods can exploit local model inform...
The paper introduces connectivity-aware client selection for decentralized federated learning, using spectral graph tools to choose a subset of structurally relevant neighbors for each client. By leveraging eigenvector embeddings and heat/diffusion kernel similarities, the method reduces communication overhead while maintaining or improving accuracy compared to uniform subsampling and full broadcasting. Experiments show up to a 3‑percentage‑point accuracy gain under strong data heterogeneity and a 50% reduction in transmitted messages.
SeedFlood is a novel decentralized fine‑tuning method for large language models that scales to billions of parameters and hundreds of clients. It leverages the seed‑reconstructible structure of zeroth‑order gradients to reduce message sizes to near‑zero, enabling efficient flooding across the network. Experiments show SeedFlood outperforms standard zeroth‑order baselines in communication efficiency and generalization, and rivals first‑order gossip methods while incurring far less communication cost.
arXiv:2607. 03171v1 Announce Type: cross Abstract: Decentralised federated learning, based on peer-to-peer communication, is increasingly proposed for on-device training of machine learning models, promising a privacy-preserving, communication-efficient training process with no risk of single-point failure.
arXiv:2607. 15053v1 Announce Type: cross Abstract: The Internet taught us that the value of a network depends on \emph{how} its nodes connect: broadcast stars scale as $V\!
arXiv:2504. 17471v2 Announce Type: replace-cross Abstract: Gossip Learning (GL) is a decentralized learning paradigm where users iteratively exchange and aggregate models with a small set of neighboring peers.
arXiv:2606. 09154v1 Announce Type: new Abstract: Decentralized SGD is a fundamental algorithm in decentralized learning, although the influence of an underlying network topology on its convergence behavior is not yet fully understood.
arXiv:2608. 06946v1 Announce Type: cross Abstract: In gossip learning, a network of nodes trains a shared model collaboratively, without a central coordinator, by repeatedly exchanging parts of their local models.
arXiv:2606. 06687v1 Announce Type: new Abstract: We investigate cluster formation, involving the number and composition of clusters, in decentralized federated learning (FL) with heterogeneous machine learning (ML) optimizers.
arXiv:2606. 28342v1 Announce Type: cross Abstract: Decentralized learning is a promising paradigm for collaborative training in mobile and pervasive systems, as it avoids a central coordinator and does not require sharing raw data.
arXiv:2606. 30499v1 Announce Type: new Abstract: Federated Learning often suffers under non-independently and identically distributed data, where a single global model may fail to represent the diversity of client distributions.
arXiv:2505. 09854v3 Announce Type: replace Abstract: As end-user device capability increases and demand for intelligent services at the Internet's edge rises, distributed learning has emerged as a key enabling technology for the intelligent edge.