?
Queuing dynamics of asynchronous Federated Learning
P. 1711–1719.
We study asynchronous federated learning mechanisms with nodes having potentially different computational speeds. In such an environment, each node is allowed to work on models with potential delays and contribute to updates to the central server at its own pace. Existing analyses of such algorithms typically depend on intractable quantities such as the maximum node delay and do not consider the underlying queuing dynamics of the system. In this paper, we propose a non-uniform sampling scheme for the central server that allows for lower delays with better complexity, taking into account the closed Jackson network structure of the associated computational graph. Our experiments clearly show a significant improvement of our method over current state-of-the-art asynchronous algorithms on image classification problems.
Beznosikov A., Kormakov G., Grigorievskiy A. et al., Journal of Optimization Theory and Applications 2026 Vol. 209 Article 18
The objective of Vertical Federated Learning (VFL) is to collectively train a model using features available on different devices while sharing the same users. This paper focuses on the saddle point reformulation of the VFL problem via the classical Lagrangian function. We first demonstrate how this formulation can be solved using deterministic methods.More importantly, we explore various stochastic modifications to ...
Added: June 17, 2026
Ali J. Dayoub, Ehab S. Suleiman, , in: Proceedings of the 2026 8th International Youth Conference on Radio Electronics, Electrical and Power Engineering (REEPE).: IEEE, 2026. Ch. 159 P. 1–5.
Abstract— Urban traffic congestion costs the global economy over $1 trillion annually, necessitating intelligent traffic signal control (ITSC) solutions. Traditional centralized approaches face critical limitations: privacy violations from vehicle trajectory data sharing, prohibitive communication overhead, and scalability challenges in heterogeneous urban environments. This paper presents a federated reinforcement learning (FRL) framework for privacy-preserving traffic signal ...
Added: April 30, 2026
Demidovich Y., Petr Ostroukhov, Malinovsky G. et al., , in: The Thirteenth International Conference on Learning Representations: ICLR 2025.: ICLR, 2025.
Non-convex Machine Learning problems typically do not adhere to the standard smoothness assumption. Based on empirical findings, Zhang et al. (2020b) proposed a more realistic generalized $(L_0,L_1)$-smoothness assumption, though it remains largely unexplored. Many existing algorithms designed for standard smooth problems need to be revised. However, in the context of Federated Learning, only a few ...
Added: July 15, 2025
Sukhoroslov O. V., Physics of Particles and Nuclei 2024 Vol. 55 P. 334–337
The paper presents DSLab, a general-purpose software framework for simulation of distributed systems. The main advantages of DSLab in comparison to other similar frameworks are versatility and extensibility, convenient and flexible programming model, high performance and ability to simulate large-scale systems. The viability of the framework is confirmed by diverse set of developed domain-specific modules ...
Added: February 22, 2025
Plassier V., Kotelevskii N., Rubashevskii A. et al., , in: Proceedings of The 27th International Conference on Artificial Intelligence and Statistics (AISTATS 2024), 2-4 May 2024, Palau de Congressos, Valencia, Spain. PMLR: Volume 238Vol. 238.: Valencia: PMLR, 2024. P. 4879–4887.
Conformal prediction (CP) stands out as a robust framework for uncertainty quantification, which is crucial for ensuring the reliability of predictions. However, common CP methods heavily rely on the data exchangeability, a condition often violated in practice. Existing approaches for tackling non-exchangeability lead to methods that are not computable beyond the simplest examples. In this ...
Added: May 30, 2024
Ali A., , in: 22nd International Conference, NEW2AN 2022, Tashkent, Uzbekistan, December 15–16, 2022, Proceedings. Internet of Things, Smart Spaces, and Next Generation Networks and Systems. LNCS, volume 13772Issue 13772.: Springer, 2023. P. 525–533.
Machine learning over distributed data collected by many clients has
important applications in use cases where data privacy is a key concern or central data storage is not an option. Federated learning has introduced solutions
for these scenarios, unlike the client-server approach, where all the training data
is centralized in the server side, the clients, in a federated ...
Added: May 18, 2023
Ali A., , in: 2022 International Conference on Smart Applications, Communications and Networking (SmartNets).: IEEE, 2022. P. 1–4.
Added: May 16, 2023
Ali A., Koucheryavy E., Ebraheem A., В кн.: Инновационные, информационные и коммуникационные технологии. Сборник трудов XIX Международной научно-практической конференции.: М.: Ассоциация выпускников и сотрудников ВВИА им. проф. Жуковского, 2022. С. 38–42.
Added: February 22, 2023
Sadiev A., Borodich E., Beznosikov A. et al., EURO Journal on Computational Optimization 2022 Vol. 10 Article 100041
This paper considers the problem of decentralized, personalized federated learning. For centralized personalized federated learning, a penalty that measures the deviation from the local model and its average, is often added to the objective function. However, in a decentralized setting this penalty is expensive in terms of communication costs, so here, a different penalty — ...
Added: October 28, 2022
Efim Luboshnikov, Makarov I., , in: Recent Trends in Analysis of Images, Social Networks and Texts. 9th International Conference, AIST 2020, Skolkovo, Moscow, Russia, October 15–16, 2020 Revised Supplementary ProceedingsVol. 12602.: Springer, 2021. Ch. 8 P. 90–101.
This article is devoted to the implementation of the federated approach to named entity recognition. The novel federated approach is designed to solve data privacy issues. The classic BiLSTM-CNNs-CRF and its modifications trained on a single machine are taken as baseline. Federated training is conducted for them. Influence of use of pretrained embedding, use of ...
Added: March 24, 2021
Российский университет дружбы народов (РУДН), 2019.
В научном издании представлены материалы XXII Международной научной конференции «Распределенные компьютерные и телекоммуникационные сети: управление, вычисление, связь» по следующим направлениям: • Алгоритмы и протоколы телекоммуникационных сетей • Управление в компьютерных и инфокоммуникационных системах • Анализ производительности, оценка QoS / QoE и эффективность сетей • Аналитическое и имитационное моделирование коммуникационных систем последующих поколений • Эволюция беспроводных ...
Added: December 25, 2019
Veretennikov A., Queueing Systems 2020 Vol. 94 No. 3 P. 243–255
A mean-field extension of the queueing system \(GI/GI/1\) is considered. The process is constructed as a Markov solution of a martingale problem. Uniqueness in distribution is also established under a slightly different set of assumptions on intensities in comparison to those required for existence. ...
Added: October 29, 2019
., 2018.
Selected Papers of the 8th International Conference "Distributed Computing and Grid-technologies in Science and Education"
Dubna, Russia, September 10-14, 2018.
Edited by
Vladimir Korenkov *
Andrey Nechaevskiy *
Tatiana Zaikina *
Elena Mazhitova *
* Joint Institute for Nuclear Research, Dubna, Moscow Region, Russia ...
Added: January 9, 2019
Toporkov V., Anna Toporkova, Yemelyanov D., , in: Intelligent Distributed Computing XIIVol. 798: Intelligent Distributed Computing XII.: Cham, Switzerland: Springer, 2018. P. 40–49.
In this work, we introduce slot selection and co-allocation algorithms for parallel jobs in distributed computing with non-dedicated and heterogeneous resources. A single slot is a time span that can be assigned to a task, which is a part of a parallel job. The job launch requires a co-allocation of a specified number of slots ...
Added: October 26, 2018
Cham, Switzerland: Springer, 2018.
This book gathers a wealth of research contributions on recent advances in intelligent and distributed computing, and which present both architectural and algorithmic findings in these fields. A major focus is placed on new techniques and applications for evolutionary computation, swarm intelligence, multi-agent systems, multi-criteria optimization and Deep/Shallow machine learning models, all of which are ...
Added: October 26, 2018
Cham, Switzerland: Springer, 2017.
This book presents the combined peer-reviewed proceedings of the tenth International Symposium on Intelligent Distributed Computing (IDC’2016), which was held in Paris, France from October 10thto 12th, 2016. The 23 contributions address a range of topics related to theory and application of intelligent distributed computing, including: Intelligent Distributed Agent-Based Systems, Ambient Intelligence and Social Networks, ...
Added: January 31, 2018
M.: RUDN, 2017.
Materials for the International Conference Analytical and Computational Methods in Probability
Theory and its Applications (ACMPT-2017)
The scientific publication presents the materials of the International Scientific Conference "Analytical and Computational Methods in Probability Theory and Its Applications" in the following main areas:
- Analytical methods in probability theory and its applications;
- Computational methods in probability theory and its ...
Added: October 18, 2017
Kulagin V., Дубинин В. Н., Информационные технологии 2016 Т. 22 № 1 С. 3–13
In this article we propose a new formalism — the structure of Petri nets (PN-structure) designed to study the models of complex systems. Structures or lattices are called partially ordered set in which any two elements there is a precise accurate lower and upper faces. In this paper we use the term "structure". Typed operations ...
Added: February 25, 2016
Toporkov V., Anna Toporkova, Tselishchev A. et al., Journal of Ambient Intelligence and Humanized Computing 2015 Vol. 6 No. 6 P. 733–740
Distributed environments with the users decoupling from resource providers are generally defined as utility Grids. This paper focuses on the problems of an efficient scheduling in virtual organizations (VOs) of utility Grids. The proposed scheduling model ensures the VO stakeholders preferences and provides strategies for effective resources utilization. An approach based on the combination of ...
Added: April 22, 2015