Low-Complexity Learning for Dynamic Spectrum Access in Multi-User Multi-Channel Networks

Kang, Sunjung; Joo, Changhee

doi:10.1109/infocom.2018.8485937

Cited by 16 publications

(8 citation statements)

References 20 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…We model each update interval as a time unit of learning and consider the MAB reward as the average cost during the interval. In particular, we use the Upper Confidence Bound (UCB) index [18,19], which is known to achieve the asymptotically optimal performance in the MAB problem, and apply it to our problem to balance the exploration-and-exploitation trade-off in finding the best threshold. We verify the performance of our learning-based γ * -threshold through simulations.…”

Section: Finding the Optimal-threshold Under Unknown Costsmentioning

confidence: 99%

Optimal Real-Time Monitoring of an Information Source Under Communication Costs

Yun

Joo

Eryılmaz

2018

2018 IEEE Conference on Decision and Control (CDC)

Self Cite

View full text Add to dashboard Cite

Section: Finding the Optimal-threshold Under Unknown Costsmentioning

confidence: 99%

Optimal Real-Time Monitoring of an Information Source Under Communication Costs

Yun

Joo

Eryılmaz

2018

2018 IEEE Conference on Decision and Control (CDC)

Self Cite

View full text Add to dashboard Cite

“…The lemma shows that the augmentation algorithm may still work well, even when the true weight is replaced with the UCB index. The proof of the lemma is analogous to Lemma A.1 of [29] but has some differences due to 'near-optimality'. It can be found in [30].…”

Section: A Regret Performance In a Single Framementioning

confidence: 99%

“…The authors of [28] have developed fully distributed schemes that can achieve the logarithmic regret without any information exchange. In [29], the authors have successfully lowered the algorithmic complexity to O(1) while achieving the logarithmic regret performance. Although these aforementioned learning algorithms are amenable to distributed implementation, they are limited to single-hop networks, such as wireless access networks, and to saturated traffic scenarios (i.e., links always have packets to send), and thus cannot accordingly respond to packet arrival dynamics.…”

Section: Introductionmentioning

confidence: 99%

A learning-based distributed algorithm for scheduling in multi-hop wireless networks

2022

Self Cite

View full text Add to dashboard Cite

We address the joint problem of learning and scheduling in multi-hop wireless network without a prior knowledge on link rates. Previous scheduling algorithms need the link rate information, and learning algorithms often require a centralized entity and polynomial complexity. These become a major obstacle to develop an efficient learning-based distributed scheme for resource allocation in large-scale multi-hop networks. In this work, by incorporating with learning algorithm, we develop provably efficient scheduling scheme under packet arrival dynamics without a priori link rate information. We extend the results to distributed implementation and evaluation their performance through simulations.

show abstract

“…In this subsection, we aim to solve P2. First, we can leverage the Lyapunov technique and transform constraint C4 into queue stability constraints [30], [31]. In detail, we introduce virtual queues 𝑄 𝑖 with the following update equation:…”

Section: B Virtual Queue and One-slot Optimization Problemmentioning

confidence: 99%

Low-Latency Federated Learning over Wireless Channels with Differential Privacy

Wei¹,

Li²,

Ma³

et al. 2021

Preprint

View full text Add to dashboard Cite

In federated learning (FL), model training is distributed over clients and local models are aggregated by a central server. The performance of uploaded models in such situations can vary widely due to imbalanced data distributions, potential demands on privacy protections, and quality of transmissions. In this paper, we aim to minimize FL training delay over wireless channels, constrained by overall training performance as well as each client's differential privacy (DP) requirement. We solve this problem in the framework of multi-agent multi-armed bandit (MAMAB) to deal with the situation where there are multiple clients confornting different unknown transmission environments, e.g., channel fading and interferences. Specifically, we first transform the long-term constraints on both training performance and each client's DP into a virtual queue based on the Lyapunov drift technique. Then, we convert the MAMAB to a max-min bipartite matching problem at each communication round, by estimating rewards with the upper confidence bound (UCB) approach. More importantly, we propose two efficient solutions to this matching problem, i.e., modified Hungarian algorithm and greedy matching with a better alternative (GMBA), in which the first one can achieve the optimal solution with a high complexity while the second one approaches a better trade-off by enabling a verified low-complexity with little performance loss. In addition, we develop an upper bound on the expected regret of this MAMAB based FL framework, which shows a linear growth over the logarithm of communication rounds, justifying its theoretical feasibility. Extensive experimental results are conducted to validate the effectiveness of our proposed algorithms, and the impacts of various parameters on the FL performance over wireless edge networks are also discussed.

show abstract

Low-Complexity Learning for Dynamic Spectrum Access in Multi-User Multi-Channel Networks

Cited by 16 publications

References 20 publications

Optimal Real-Time Monitoring of an Information Source Under Communication Costs

Optimal Real-Time Monitoring of an Information Source Under Communication Costs

A learning-based distributed algorithm for scheduling in multi-hop wireless networks

Low-Latency Federated Learning over Wireless Channels with Differential Privacy

Contact Info

Product

Resources

About