A Case of Distributed Optimization in Adversarial Environment

Ravi, Nikhil; Scaglione, Anna; Nedić, Angelia

doi:10.1109/icassp.2019.8683442

Cited by 20 publications

(14 citation statements)

References 19 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Ravi et al [88] analyzed possible behavior of malicious agents in the system. Suppose the malicious agents intend to manipulate its objective function, such that the output using cost functions from all agents x a will deviate from a correct output x * by a vector , i.e.…”

Section: Alternative Adversarial Modelsmentioning

confidence: 99%

A Survey on Fault-tolerance in Distributed Optimization and Machine Learning

Liu

2021

Preprint

View full text Add to dashboard Cite

The robustness of distributed optimization is an emerging field of study, motivated by various applications of distributed optimization including distributed machine learning, distributed sensing, and swarm robotics. With the rapid expansion of the scale of distributed systems, resilient distributed algorithms for optimization are needed, in order to mitigate system failures, communication issues, or even malicious attacks. This survey investigates the current state of fault-tolerance research in distributed optimization, and aims to provide an overview of the existing studies on both fault-tolerant distributed optimization theories and applicable algorithms.

show abstract

Section: Alternative Adversarial Modelsmentioning

confidence: 99%

A Survey on Fault-tolerance in Distributed Optimization and Machine Learning

Liu

2021

Preprint

View full text Add to dashboard Cite

show abstract

“…Note that all the above‐mentioned algorithms are proposed under the assumption that the communication network surrounding is benign. In fact, complex cyber‐physical networks are usually in open and hostile environments and thus suffering from cyber or physical attacks inevitably 33‐42 . Distributed optimization problems in the adversarial environment also come into the view of researchers, where partial nodes are adversarial or faulty.…”

Section: Introductionmentioning

confidence: 99%

“…Distributed optimization problems in the adversarial environment also come into the view of researchers, where partial nodes are adversarial or faulty. In the case that malicious nodes intend to drive the solutions of normal nodes away from the global minimizer, Ravi et al 36 use the trends of averaged gradient variations to detect the neighbors that are likely to be malicious, by observing that the gradient of attackers tends to be large. A fault‐tolerant optimal iterative distributed algorithm 37 is proposed under a completely connected network with at most a third of the nodes being faulty, where only two neighbors' estimations and another two neighbors' gradients are used for each node's update.…”

Section: Introductionmentioning

confidence: 99%

Resilient consensus‐based distributed optimization under deception attacks

Qin

et al. 2020

Intl J Robust & Nonlinear

View full text Add to dashboard Cite

Summary This article considers distributed optimization problems of complex cyber‐physical networks, whose goal is to minimize a global function consisting of a sum of local functions possessed by each node, when the communication network is suffering L‐local deception attacks. After showing that merely 1‐local deception attacks can arbitrarily affect the outcome of any distributed optimization algorithms without being detected, we propose a resilient consensus‐based distributed optimization algorithm, where the estimation for the optimizer of each node is updated according to its subgradient and its partial neighbors' estimation. Then, we provide the conditions for the proposed algorithm to ensure that all the nodes can make an agreement and converge to the convex hull of the local optimizer of their functions in the presence of L‐local deception attacks. Finally, some simulation examples are presented to demonstrate the effectiveness of the proposed algorithm.

show abstract

“…A handful of recent papers have considered this problem for the case where agent misbehavior follows prescribed patterns [6], [7]. A more general (and serious) form of misbehavior is captured by the Byzantine adversary model from computer science, where misbehaving agents can send arbitrary (and conflicting) values to their neighbors at each iteration of the algorithm.…”

Section: Introductionmentioning

confidence: 99%

Byzantine-Resilient Distributed Optimization of Multi-Dimensional Functions

Kuwaranancharoen¹,

Lei²,

Sundaram³

2020

Preprint

View full text Add to dashboard Cite

The problem of distributed optimization requires a group of agents to reach agreement on a parameter that minimizes the average of their local cost functions using information received from their neighbors. While there are a variety of distributed optimization algorithms that can solve this problem, they are typically vulnerable to malicious (or "Byzantine") agents that do not follow the algorithm. Recent attempts to address this issue focus on single dimensional functions, or provide analysis under certain assumptions on the statistical properties of the functions at the agents. In this paper, we propose a resilient distributed optimization algorithm for multidimensional convex functions. Our scheme involves two filtering steps at each iteration of the algorithm: (1) distance-based and (2) component-wise removal of extreme states. We show that this algorithm can mitigate the impact of up to F Byzantine agents in the neighborhood of each regular node, without knowing the identities of the Byzantine agents in advance. In particular, we show that if the network topology satisfies certain conditions, all of the regular states are guaranteed to asymptotically converge to a bounded region that contains the global minimizer.

show abstract

A Case of Distributed Optimization in Adversarial Environment

Cited by 20 publications

References 19 publications

A Survey on Fault-tolerance in Distributed Optimization and Machine Learning

A Survey on Fault-tolerance in Distributed Optimization and Machine Learning

Resilient consensus‐based distributed optimization under deception attacks

Byzantine-Resilient Distributed Optimization of Multi-Dimensional Functions

Contact Info

Product

Resources

About