Distributed dual averaging for convex optimization under communication delays

By Konstantinos I. Tsianos, Michael G. Rabbat. 2012

115 citations recorded by the source corpus.

Subfields: Computer Networks and Communications

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Computer Networks and Communications

1. Ali H. Sayed
2. Angelia Nedić
3. Kun Yuan
4. Wotao Yin
5. Qing Ling
6. Aryan Mokhtari
7. Michael G. Rabbat
8. Ji Liu

Citations in the corpus

• Central Server Free Federated Learning over Single-sided Trust Social Networks (2019, background): “Such a requirement is quite restrictive, because not all networks admit a doubly stochastic matrix (Gharesifard and Cortés 2010), and relinquishing double stochasticity can introduce bias in optimization (Ram et al., 2010; Tsianos and Rabbat, 2012).”
• Push-Sum Distributed Dual Averaging for convex optimization (2012, methodology): “PUSH-SUM CONSENSUS In [8] it is proven that restricting to doubly stochastic consensus protocols in distributed dual averaging is not necessary and it is still possible to converge to the optimum with a general row stochastic protocol P .”
• Asynchronous Saddle Point Algorithm for Stochastic Optimization in Heterogeneous Networks (2019, methodology): “Techniques that are good for distributed convex optimization, for example, those based on penalty method [11], [12] or on Lagrange duality [13], [14], have in most cases translated into the distributed stochastic domain without major hurdles, as in [12], [15], [16].”
• Consensus-based distributed optimization: Practical issues and applications in large-scale machine learning (2012, background): “For example, [21] shows that a simple reweighing of the objective removes the bias and achieves averaging for any row-stochastic matrix P with stationary distribution π (see also the scaled agreement algorithm in [22]):”
• Decentralized Consensus Optimization With Asynchrony and Delays (2016, background): “We also distinguish our setting from the fixed communication delay setting [30], [31], where the information passing through each edge takes a fixed number of iterations to arrive.”
• Communication/Computation Tradeoffs in Consensus-Based Distributed Optimization (2012, background): “In addition, they combine simplicity of implementation with robustness to node failures and are resilient to communication delays [9].”
• Decentralized consensus optimization with asynchrony and delays (2016, background): “We also distinguish our setting from the fixed communication delay setting [27], [28], where the information passing through each edge takes a fixed number of iterations to arrive.”