Enhancing Geometric Deep Learning via Graph Filter Deconvolution

Yang, Jingkang; Segarra, Santiago

doi:10.1109/globalsip.2018.8646544

Cited by 13 publications

(8 citation statements)

References 30 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…This has led to several approaches that tried to adjust CNNs to the wireless setting [15], [37], [38] capitalizing on, e.g., the spatial disposition of wireless networks [39]. We adopt an alternative direction [25], [26], [40], where graph neural networks (GNNs) [41]- [47] are used to parameterize the power allocation function, thus leveraging the natural representation of wireless networks as graphs. GNNs utilize the structural relationships between nodes to locally process instantaneous channel state information.…”

mentioning

confidence: 99%

Unfolding WMMSE using Graph Neural Networks for Efficient Power Allocation

Chowdhury¹,

Verma²,

Rao³

et al. 2020

Preprint

Self Cite

View full text Add to dashboard Cite

We study the problem of optimal power allocation in a single-hop ad hoc wireless network. In solving this problem, we depart from classical purely model-based approaches and propose a hybrid method that retains key modeling elements in conjunction with data-driven components. More precisely, we put forth a neural network architecture inspired by the algorithmic unfolding of the iterative weighted minimum mean squared error (WMMSE) method, that we denote by unfolded WMMSE (UWMMSE). The learnable weights within UWMMSE are parameterized using graph neural networks (GNNs), where the time-varying underlying graphs are given by the fading interference coefficients in the wireless network. These GNNs are trained through a gradient descent approach based on multiple instances of the power allocation problem. We show that the proposed architecture is permutation equivariant, thus facilitating generalizability across network topologies. Comprehensive numerical experiments illustrate the performance attained by UWMMSE along with its robustness to hyper-parameter selection and generalizability to unseen scenarios such as different network densities and network sizes.Index terms-WMMSE, power allocation, graph neural networks, deep learning, algorithm unfolding. I. INTRODUCTIONPower and bandwidth are fundamental resources in communication, playing a key role in determining the effective capacity of a wireless channel [1], [2]. In modern wireless communication systems, the scope of resources has been broadened to include beams in a multiple-input-multipleoutput (MIMO) system, time slots in a time-division multiple access system (TDMA), frequency sub-bands in a frequencydivision multiple access (FDMA) system, spreading codes in a code-division multiple access (CDMA) system, among several others [3]. Optimal allocation of these resources under randomly varying channel characteristics and user demands is essential for the smooth operation of wireless systems. In particular, power allocation in a wireless ad hoc network is crucial to mitigate multi-user interference, one of the main performance-limiting factors. In addition, transmission power of a mobile user is in itself a scarce resource. Indeed, careful

show abstract

mentioning

confidence: 99%

Unfolding WMMSE using Graph Neural Networks for Efficient Power Allocation

Chowdhury¹,

Verma²,

Rao³

et al. 2020

Preprint

Self Cite

View full text Add to dashboard Cite

show abstract

“…This has led to several approaches that tried to adjust CNNs to the wireless setting [28], [35]- [37]. Here, we adopt an alternative direction [33], [38], in which graph neural networks (GNNs) [39]- [42] are used to incorporate the topology of the wireless network into the learning algorithm.…”

Section: Introductionmentioning

confidence: 99%

Link Scheduling using Graph Neural Networks

Zhao¹,

Verma²,

Rao³

et al. 2021

Preprint

Self Cite

View full text Add to dashboard Cite

Efficient scheduling of transmissions is a key problem in wireless networks. The main challenge stems from the fact that optimal link scheduling involves solving a maximum weighted independent set (MWIS) problem, which is known to be NP-hard. For practical link scheduling schemes, centralized and distributed greedy heuristics are commonly used to approximate the solution to the MWIS problem. However, these greedy schemes mostly ignore important topological information of the wireless network. To overcome this limitation, we propose fast heuristics based on graph convolutional networks (GCNs) that can be implemented in centralized and distributed manners. Our centralized MWIS solver is based on tree search guided by a trainable GCN module and 1-step rollout. In our distributed MWIS solver, a trainable GCN module learns topology-aware node embeddings that are combined with the network weights before calling a distributed greedy solver. Test results on mediumsized wireless networks show that a GCN-based centralized MWIS solver can reach a near-optimal solution quickly. Moreover, we demonstrate that a shallow GCN-based distributed MWIS scheduler can reduce by nearly half the suboptimality gap of the distributed greedy solver with minimal increase in complexity. The proposed scheduling solutions also exhibit good generalizability across graph and weight distributions.

show abstract

“…This has led to several approaches that tried to adjust CNNs to the wireless setting [15,[21][22][23]. Here, we adopt an alternative direction [20,24], where graph neural networks (GNNs) [25][26][27][28] are used to incorporate the topology of the wireless network in the learning algorithm. Our approach is also in line with the recent trend of using deep learning to find approximate solutions to combinatorial problems on graphs [29,30].…”

Section: Introductionmentioning

confidence: 99%

Distributed Scheduling Using Graph Neural Networks

Zhao

Verma²,

Rao³

et al. 2021

ICASSP 2021 - 2021 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)

Self Cite

View full text Add to dashboard Cite

A fundamental problem in the design of wireless networks is to efficiently schedule transmission in a distributed manner. The main challenge stems from the fact that optimal link scheduling involves solving a maximum weighted independent set (MWIS) problem, which is NP-hard. For practical link scheduling schemes, distributed greedy approaches are commonly used to approximate the solution of the MWIS problem. However, these greedy schemes mostly ignore important topological information of the wireless networks. To overcome this limitation, we propose a distributed MWIS solver based on graph convolutional networks (GCNs). In a nutshell, a trainable GCN module learns topology-aware node embeddings that are combined with the network weights before calling a greedy solver. In small-to middle-sized wireless networks with tens of links, even a shallow GCN-based MWIS scheduler can leverage the topological information of the graph to reduce in half the suboptimality gap of the distributed greedy solver with good generalizability across graphs and minimal increase in complexity.

show abstract

Enhancing Geometric Deep Learning via Graph Filter Deconvolution

Cited by 13 publications

References 30 publications

Unfolding WMMSE using Graph Neural Networks for Efficient Power Allocation

Unfolding WMMSE using Graph Neural Networks for Efficient Power Allocation

Link Scheduling using Graph Neural Networks

Distributed Scheduling Using Graph Neural Networks

Contact Info

Product

Resources

About