arXiv subjects

Osman Yagan

Publications and source records attributed to Osman Yagan.

At least 19 recordsLinked to original sources

arXiv2026

TANGCO: Learning Topology-Aware Capacity Allocation for Overload-driven Cascading Failures

Networked systems, from power grids to traffic networks and cloud clusters, carry loads across nodes with limited capacity. A node whose load exceeds its capacity fails and sheds its load onto its neighbors, which can trigger a system-wide cascade. We study how to allocate a fixed capacity budget across nodes to resist these cascades under local load redistribution. The problem is difficult because no optimal allocation is known, and the fail-or-survive objective is non-differentiable and piecewise constant, so exact and gradient-based optimization methods do not directly apply. We introduce TANGCO (Topology-Aware Neural Graph-Guided Capacity Optimization), which uses a graph neural network policy trained through the cascade simulator with policy-gradient learning and a heuristic anchor. We evaluate TANGCO on five synthetic graph families and five real networks spanning power, road, air, and Internet topologies. The learned policy improves on the best of four hand-designed heuristics in all 450 synthetic instances and in 40 of 45 real-network conditions, with robustness gains ranging from 1.6% to 246%. The learned policies transfer to unseen graphs within a family and partially across related topologies, and TANGCO$^{pre}$, pre-trained on synthetic graphs, matches per-network training on unseen real networks. Training scales near-linearly with graph size, and TANGCO$^{pre}$ allocates on a new network with no per-target training, matching the deployment cost of a hand-designed heuristic. Free-vector variants without the GNN, stay close to the heuristics, so the graph representation carries the gain beyond numerical search. Finally, analysis of the learned allocations identifies when local risk is sufficient, leads to an improved closed-form heuristic, and reveals the regimes where a topology-aware learned policy remains necessary.

cs.LG↗

arXiv2025

Achieving Logarithmic Regret in KL-Regularized Zero-Sum Markov Games

Reverse Kullback-Leibler (KL) divergence-based regularization with respect to a fixed reference policy is widely used in modern reinforcement learning to preserve the desired traits of the reference policy and sometimes to promote exploration (using uniform reference policy, known as entropy regularization). Beyond serving as a mere anchor, the reference policy can also be interpreted as encoding prior knowledge about good actions in the environment. In the context of alignment, recent game-theoretic approaches have leveraged KL regularization with pretrained language models as reference policies, achieving notable empirical success in self-play methods. Despite these advances, the theoretical benefits of KL regularization in game-theoretic settings remain poorly understood. In this work, we develop and analyze algorithms that provably achieve improved sample efficiency under KL regularization. We study both two-player zero-sum matrix games and Markov games: for matrix games, we propose OMG, an algorithm based on best response sampling with optimistic bonuses, and extend this idea to Markov games through the algorithm SOMG, which also uses best response sampling and a novel concept of superoptimistic bonuses. Both algorithms achieve a logarithmic regret in $T$ that scales inversely with the KL regularization strength $\beta$ in addition to the traditional $\widetilde{\mathcal{O}}(\sqrt{T})$ regret without the $\beta^{-1}$ dependence.

cs.LG↗

arXiv2025

On Balancing Sparsity with Reliable Connectivity in Distributed Network Design with Random K-out Graphs

In several applications in distributed systems, an important design criterion is ensuring that the network is sparse, i.e., does not contain too many edges, while achieving reliable connectivity. Sparsity ensures communication overhead remains low, while reliable connectivity is tied to reliable communication and inference on decentralized data reservoirs and computational resources. A class of network models called random K-out graphs appear widely as a heuristic to balance connectivity and sparsity, especially in settings with limited trust, e.g., privacy-preserving aggregation of networked data in which networks are deployed. However, several questions remain regarding how to choose network parameters in response to different operational requirements, including the need to go beyond asymptotic results and the ability to model the stochastic and adversarial environments. To address this gap, we present theorems to inform the choice of network parameters that guarantee reliable connectivity in regimes where nodes can be finite or unreliable. We first derive upper and lower bounds for probability of connectivity in random K-out graphs when the number of nodes is finite. Next, we analyze the property of r-robustness, a stronger notion than connectivity that enables resilient consensus in the presence of malicious nodes. Finally, motivated by aggregation mechanisms based on pairwise masking, we model and analyze the impact of a subset of adversarial nodes, modeled as deletions, on connectivity and giant component size - metrics that are closely tied to privacy guarantees. Together, our results pave the way for end-to-end performance guarantees for a suite of algorithms for reliable inference on networks.

cs.SI↗

arXiv2025

Cost-aware LLM-based Online Dataset Annotation

Recent advances in large language models (LLMs) have enabled automated dataset labeling with minimal human supervision. While majority voting across multiple LLMs can improve label reliability by mitigating individual model biases, it incurs high computational costs due to repeated querying. In this work, we propose a novel online framework, Cost-aware Majority Voting (CaMVo), for efficient and accurate LLM-based dataset annotation. CaMVo adaptively selects a subset of LLMs for each data instance based on contextual embeddings, balancing confidence and cost without requiring pre-training or ground-truth labels. Leveraging a LinUCB-based selection mechanism and a Bayesian estimator over confidence scores, CaMVo estimates a lower bound on labeling accuracy for each LLM and aggregates responses through weighted majority voting. Our empirical evaluation on the MMLU and IMDB Movie Review datasets demonstrates that CaMVo achieves comparable or superior accuracy to full majority voting while significantly reducing labeling costs. This establishes CaMVo as a practical and robust solution for cost-efficient annotation in dynamic labeling environments.

cs.LG↗

arXiv2025

Bandits with Anytime Knapsacks

We consider bandits with anytime knapsacks (BwAK), a novel version of the BwK problem where there is an \textit{anytime} cost constraint instead of a total cost budget. This problem setting introduces additional complexities as it mandates adherence to the constraint throughout the decision-making process. We propose SUAK, an algorithm that utilizes upper confidence bounds to identify the optimal mixture of arms while maintaining a balance between exploration and exploitation. SUAK is an adaptive algorithm that strategically utilizes the available budget in each round in the decision-making process and skips a round when it is possible to violate the anytime cost constraint. In particular, SUAK slightly under-utilizes the available cost budget to reduce the need for skipping rounds. We show that SUAK attains the same problem-dependent regret upper bound of $ O(K \log T)$ established in prior work under the simpler BwK framework. Finally, we provide simulations to verify the utility of SUAK in practical settings.

cs.LG↗

arXiv2024

FedSPD: A Soft-clustering Approach for Personalized Decentralized Federated Learning

Federated learning has recently gained popularity as a framework for distributed clients to collaboratively train a machine learning model using local data. While traditional federated learning relies on a central server for model aggregation, recent advancements adopt a decentralized framework, enabling direct model exchange between clients and eliminating the single point of failure. However, existing decentralized frameworks often assume all clients train a shared model. Personalizing each client's model can enhance performance, especially with heterogeneous client data distributions. We propose FedSPD, an efficient personalized federated learning algorithm for the decentralized setting, and show that it learns accurate models even in low-connectivity networks. To provide theoretical guarantees on convergence, we introduce a clustering-based framework that enables consensus on models for distinct data clusters while personalizing to unique mixtures of these clusters at different clients. This flexibility, allowing selective model updates based on data distribution, substantially reduces communication costs compared to prior work on personalized federated learning in decentralized settings. Experimental results on real-world datasets show that FedSPD outperforms multiple decentralized variants of personalized federated learning algorithms, especially in scenarios with low-connectivity networks.

cs.LG↗

arXiv2024

On the Interplay of Clustering and Evolution in the Emergence of Epidemic Outbreaks

In an increasingly interconnected world, a key scientific challenge is to examine mechanisms that lead to the widespread propagation of contagions, such as misinformation and pathogens, and identify risk factors that can trigger large-scale outbreaks. Underlying both the spread of disease and misinformation epidemics is the evolution of the contagion as it propagates, leading to the emergence of different strains, e.g., through genetic mutations in pathogens and alterations in the information content. Recent studies have revealed that models that do not account for heterogeneity in transmission risks associated with different strains of the circulating contagion can lead to inaccurate predictions. However, existing results on multi-strain spreading assume that the network has a vanishingly small clustering coefficient, whereas clustering is widely known to be a fundamental property of real-world social networks. In this work, we investigate spreading processes that entail evolutionary adaptations on random graphs with tunable clustering and arbitrary degree distributions. We derive a mathematical framework to quantify the epidemic characteristics of a contagion that evolves as it spreads, with the structure of the underlying network as given via arbitrary {\em joint} degree distributions of single-edges and triangles. To the best of our knowledge, our work is the first to jointly analyze the impact of clustering and evolution on the emergence of epidemic outbreaks. We supplement our theoretical finding with numerical simulations and case studies, shedding light on the impact of clustering on contagion spread.

cs.SI↗

arXiv2023

The Role of Masks in Mitigating Viral Spread on Networks

Masks have remained an important mitigation strategy in the fight against COVID-19 due to their ability to prevent the transmission of respiratory droplets between individuals. In this work, we provide a comprehensive quantitative analysis of the impact of mask-wearing. To this end, we propose a novel agent-based model of viral spread on networks where agents may either wear no mask or wear one of several types of masks with different properties (e.g., cloth or surgical). We derive analytical expressions for three key epidemiological quantities: the probability of emergence, the epidemic threshold, and the expected epidemic size. In particular, we show how the aforementioned quantities depend on the structure of the contact network, viral transmission dynamics, and the distribution of the different types of masks within the population. Through extensive simulations, we then investigate the impact of different allocations of masks within the population and trade-offs between the outward efficiency and inward efficiency of the masks. Interestingly, we find that masks with high outward efficiency and low inward efficiency are most useful for controlling the spread in the early stages of an epidemic, while masks with high inward efficiency but low outward efficiency are most useful in reducing the size of an already large spread. Lastly, we study whether degree-based mask allocation is more effective in reducing the probability of epidemic as well as epidemic size compared to random allocation. The result echoes the previous findings that mitigation strategies should differ based on the stage of the spreading process, focusing on source control before the epidemic emerges and on self-protection after the emergence.

cs.SI↗

arXiv2023

Spreading Processes with Mutations over Multi-layer Networks

A key scientific challenge during the outbreak of novel infectious diseases is to predict how the course of the epidemic changes under different countermeasures that limit interaction in the population. Most epidemiological models do not consider the role of mutations and heterogeneity in the type of contact events. However, pathogens have the capacity to mutate in response to changing environments, especially caused by the increase in population immunity to existing strains and the emergence of new pathogen strains poses a continued threat to public health. Further, in light of differing transmission risks in different congregate settings (e.g., schools and offices), different mitigation strategies may need to be adopted to control the spread of infection. We analyze a multi-layer multi-strain model by simultaneously accounting for i) pathways for mutations in the pathogen leading to the emergence of new pathogen strains, and ii) differing transmission risks in different congregate settings, modeled as network-layers. Assuming complete cross-immunity among strains, namely, recovery from any infection prevents infection with any other (an assumption that will need to be relaxed to deal with COVID-19 or influenza), we derive the key epidemiological parameters for the proposed multi-layer multi-strain framework. We demonstrate that reductions to existing network-based models that discount heterogeneity in either the strain or the network layers can lead to incorrect predictions for the course of the outbreak. In addition, our results highlight that the impact of imposing/lifting mitigation measures concerning different contact network layers (e.g., school closures or work-from-home policies) should be evaluated in connection with their effect on the likelihood of the emergence of new pathogen strains.

physics.soc-ph↗

arXiv2022

Spreading processes with population heterogeneity over multi-layer networks

It's been controversial whether re-opening school will facilitate viral spread among household communities with mitigation strategies such as mask-wearing in place. In this work, we propose an epidemiological model that explores the viral transmission over the multi-layer contact network composed of the school layer and community layer with population heterogeneity on mask-wearing behavior. We derive analytical expressions for three key epidemiological quantities: the probability of emergence, the epidemic threshold, and the expected epidemic size. In particular, we show how the aforementioned quantities depend on the structure of the multi-layer contact network, viral transmission dynamics, and the distribution of the different types of masks within the population. Through extensive simulations, our analytical results show near-perfect agreement with the simulation results with a limited number of nodes. Utilizing the model, we study the impact of the opening/closure of the school layer on the viral transmission dynamics with various mask-wearing scenarios. Interestingly, we found that it's safe to open the school layer with the proper proportion of good-quality masks in the population. Moreover, we validate the theory of the trade-off between source-control and self-protection over a single layer by Tian et al on our multi-layer setting. We conclude that even on a multi-layer network, it's of great significance to treat the spreading process as two distinct phases in mind when considering mitigation strategies. Besides, we would like to remark that our model of spreading process over multi-layer networks with population heterogeneity can also be applied to various other domains, such as misinformation control.

cs.SI↗

arXiv2022

Dynamic Coupling Strategy for Interdependent Network Systems Against Cascading Failures

Cascading failures are a common phenomenon in complex networked systems where failures at only a few nodes may trigger a process of sequential failure. We applied a flow redistribution model to investigate the robustness against cascading failures in modern systems carrying flows/loads (i.e. power grid, transportation system, etc.) that contain multiple interdependent networks. In such a system, the coupling coefficients between networks, which determine how much flows/loads are redistributed between networks, are a key factor determining the robustness to cascading failures. We derive recursive expressions to characterize the evolution of such a system under dynamic network coupling. Using these expressions, we enhance the robustness of interdependent network systems by dynamically adjusting the coupling coefficients based on current system situations, minimizing the subsequent failures. The analytical and simulation results show a significant improvement in robustness compared to prior work, which considers only fixed coupling coefficients. Our proposed Step-wise Optimization (SWO) method not only shows good performance against cascading failures, but also offers better computational complexity, scalability to multiple networks, and flexibility to different attack types. We show in simulation that SWO provides robustness against cascading failures for multiple different network topologies.

cs.NI↗

arXiv2021

Existence and Size of the Giant Component in Inhomogeneous Random K-out Graphs

Random K-out graphs are receiving attention as a model to construct sparse yet well-connected topologies in distributed systems including sensor networks, federated learning, and cryptocurrency networks. In response to the growing heterogeneity in emerging real-world networks, where nodes differ in resources and requirements, inhomogeneous random K-out graphs, denoted by $H(n;μ,K_n)$, were proposed recently. Motivated by practical settings where establishing links is costly and only a bounded choice of $K_n$ is feasible ($K_n = O(1)$), we study the size of the largest connected sub-network of $H(n;μ,K_n)$, We first show that the trivial condition of $K_n \geq 2$ for all $n$ is sufficient to ensure that $H(n;μ,K_n)$, contains a giant component of size $n-O(1)$ whp. Next, to model settings where nodes can fail or get compromised, we investigate the size of the largest connected sub-network in $H(n;μ,K_n)$, when $d_n$ nodes are selected uniformly at random and removed from the network. We show that if $d_n=O(1)$, a giant component of size $n- \OO(1)$ persists for all $K_n \geq 2$ whp. Further, when $d_n=o(n)$ nodes are removed from $H(n;μ,K_n)$, the remaining nodes contain a giant component of size $n(1-o(1))$ whp for all $K_n \geq 2$. We present numerical results to demonstrate the size of the largest connected component when the number of nodes is finite.

cs.IT↗

arXiv2021

On the Connectivity and Giant Component Size of Random K-out Graphs Under Randomly Deleted Nodes

Random K-out graphs, denoted $\mathbb{H}(n;K)$, are generated by each of the $n$ nodes drawing $K$ out-edges towards $K$ distinct nodes selected uniformly at random, and then ignoring the orientation of the arcs. Recently, random K-out graphs have been used in applications as diverse as random (pairwise) key predistribution in ad-hoc networks, anonymous message routing in crypto-currency networks, and differentially-private federated averaging. In many applications, connectivity of the random K-out graph when some of its nodes are dishonest, have failed, or have been captured is of practical interest. We provide a comprehensive set of results on the connectivity and giant component size of $\mathbb{H}(n;K_n,γ_n)$, i.e., random K-out graph when $γ_n$ of its nodes, selected uniformly at random, are deleted. First, we derive conditions for $K_n$ and $n$ that ensure, with high probability (whp), the connectivity of the remaining graph when the number of deleted nodes is $γ_n=Ω(n)$ and $γ_n=o(n)$, respectively. Next, we derive conditions for $\mathbb{H}(n;K_n,γ_n)$ to have a giant component, i.e., a connected subgraph with $Ω(n)$ nodes, whp. This is also done for different scalings of $γ_n$ and upper bounds are provided for the number of nodes outside the giant component. Simulation results are presented to validate the usefulness of the results in the finite node regime.

cs.IT↗

arXiv2020

IRS-Assisted Millimeter Wave Communications: Joint Power Allocation and Beamforming Design

Intelligent reflecting surface (IRS) technology offers more feasible propagation paths for millimeter-wave (mmWave) communication systems to overcome blockage than existing technologies. In this paper, we consider a downlink wireless system with the IRS and formulate a joint power allocation and beamforming design problem to maximize the weighted sum-rate, which is a multi-variable optimization problem. To solve the problem, we propose a novel alternating manifold optimization based beamforming algorithm. Simulation results show that our proposed optimization algorithm outperforms existing algorithms significantly in improving the weighted sum-rate of the wireless communication system.

eess.SP↗

arXiv2020

Analysis of the Impact of Mask-wearing in Viral Spread: Implications for COVID-19

Masks are used as part of a comprehensive strategy of measures to limit transmission and save lives during the COVID-19 pandemic. Research about the impact of mask-wearing in the COVID-19 pandemic has raised formidable interest across multiple disciplines. In this paper, we investigate the impact of mask-wearing in spreading processes over complex networks. This is done by studying a heterogeneous bond percolation process over a multi-type network model, where nodes can be one of two types (mask-wearing, and not-mask-wearing). We provide analytical results that accurately predict the expected epidemic size and probability of emergence as functions of the characteristics of the spreading process (e.g., transmission probabilities, inward and outward efficiency of the masks, etc.), the proportion of mask-wearers in the population, and the structure of the underlying contact network. In addition to the theoretical analysis, we also conduct extensive simulations on random networks. We also comment on the analogy between the mask-model studied here and the multiple-strain viral spreading model with mutations studied recently by Eletreby et al.

cs.SI↗

arXiv2020

Tight Bounds for Connectivity of Random K-out Graphs

Random K-out graphs are used in several applications including modeling by sensor networks secured by the random pairwise key predistribution scheme, and payment channel networks. The random K-out graph with $n$ nodes is constructed as follows. Each node draws an edge towards $K$ distinct nodes selected uniformly at random. The orientation of the edges is then ignored, yielding an undirected graph. An interesting property of random K-out graphs is that they are connected almost surely in the limit of large $n$ for any $K \geq2$. This means that they attain the property of being connected very easily, i.e., with far fewer edges ($O(n)$) as compared to classical random graph models including Erdős-Rényi graphs ($O(n \log n)$). This work aims to reveal to what extent the asymptotic behavior of random K-out graphs being connected easily extends to cases where the number $n$ of nodes is small. We establish upper and lower bounds on the probability of connectivity when $n$ is finite. Our lower bounds improve significantly upon the existing results, and indicate that random K-out graphs can attain a given probability of connectivity at much smaller network sizes than previously known. We also show that the established upper and lower bounds match order-wise; i.e., further improvement on the order of $n$ in the lower bound is not possible. In particular, we prove that the probability of connectivity is $1-Θ({1}/{n^{K^2-1}})$ for all $K \geq 2$. Through numerical simulations, we show that our bounds closely mirror the empirically observed probability of connectivity.

cs.IT↗

arXiv2019

On Connectivity and Robustness in Random Intersection Graphs

Random intersection graphs have received much attention recently and been used in a wide range of applications ranging from key predistribution in wireless sensor networks to modeling social networks. For these graphs, each node is equipped with a set of objects in a random manner, and two nodes have an undirected edge in between if they have at least one object in common. In this paper, we investigate connectivity and robustness in a general random intersection graph model. Specifically, we establish sharp asymptotic zero-one laws for k-connectivity and k-robustness, as well as the asymptotically exact probability of k-connectivity, for any positive integer k. The k-connectivity property quantifies how resilient is the connectivity of a graph against node or edge failures, while k-robustness measures the effectiveness of local-information-based consensus algorithms (that do not use global graph topology information) in the presence of adversarial nodes. In addition to presenting the results under the general random intersection graph model, we consider two special cases of the general model, a binomial random intersection graph and a uniform random intersection graph, which both have numerous applications as well. For these two specialized graphs, our results on asymptotically exact probabilities of k-connectivity and asymptotic zero-one laws for k-robustness are also novel in the literature.

cs.DM↗

arXiv2019

k-Connectivity in Random Key Graphs with Unreliable Links

Random key graphs form a class of random intersection graphs and are naturally induced by the random key predistribution scheme of Eschenauer and Gligor for securing wireless sensor network (WSN) communications. Random key graphs have received much interest recently, owing in part to their wide applicability in various domains including recommender systems, social networks, secure sensor networks, clustering and classification analysis, and cryptanalysis to name a few. In this paper, we study connectivity properties of random key graphs in the presence of unreliable links. Unreliability of the edges are captured by independent Bernoulli random variables, rendering edges of the graph to be on or off independently from each other. The resulting model is an intersection of a random key graph and an Erdos-Renyi graph, and is expected to be useful in capturing various real-world networks; e.g., with secure WSN applications in mind, link unreliability can be attributed to harsh environmental conditions severely impairing transmissions. We present conditions on how to scale this model's parameters so that i) the minimum node degree in the graph is at least k, and ii) the graph is k-connected, both with high probability as the number of nodes becomes large. The results are given in the form of zeroone laws with critical thresholds identified and shown to coincide for both graph properties. These findings improve the previous results by Rybarczyk on the k-connectivity of random key graphs (with reliable links), as well as the zero-one laws by Yagan on the 1-connectivity of random key graphs with unreliable links.

cs.IT↗