Learning to Optimize: Training Deep Neural Networks for Interference Management

Sun, Haoran; Chen, Xiangyi; Shi, Qingjiang; Mei, Hong; Fu, Xiao; Sidiropoulos, Nicholas D.

doi:10.1109/tsp.2018.2866382

Cited by 725 publications

(607 citation statements)

References 16 publications

Supporting

Mentioning

606

Contrasting

Order By: Relevance

“…II illustrates MLP and CNN based approaches for power control. From the numerical experiments in [1], [3], we observe a performance loss when K gets larger. For example, in [1], the performance gap to the WMMSE algorithm is 3% when K = 10 and becomes 12% when K = 30.…”

Section: B Existing Approaches' Limitationsmentioning

confidence: 87%

“…From the numerical experiments in [1], [3], we observe a performance loss when K gets larger. For example, in [1], the performance gap to the WMMSE algorithm is 3% when K = 10 and becomes 12% when K = 30. From the perspective of approximation theory, an MLP with a sufficient number of parameters can learn anything if we have sufficient training samples [15].…”

Section: B Existing Approaches' Limitationsmentioning

confidence: 87%

See 1 more Smart Citation

A Graph Neural Network Approach for Scalable Wireless Power Control

Shen

Shi

Zhang

et al. 2019

2019 IEEE Globecom Workshops (GC Wkshps)

View full text Add to dashboard Cite

Deep neural networks have recently emerged as a disruptive technology to solve NP-hard wireless resource allocation problems in a real-time manner. However, the adopted neural network structures, e.g., multi-layer perceptron (MLP) and convolutional neural network (CNN), are inherited from deep learning for image processing tasks, and thus are not tailored to problems in wireless networks. In particular, the performance of these methods deteriorates dramatically when the wireless network size becomes large. In this paper, we propose to utilize graph neural networks (GNNs) to develop scalable methods for solving the power control problem in K-user interference channels. Specifically, a K-user interference channel is first modeled as a complete graph, where the quantitative information of wireless channels is incorporated as the features of the graph. We then propose an interference graph convolutional neural network (IGCNet) to learn the optimal power control in an unsupervised manner. It is shown that one-layer IGCNet is a universal approximator to continuous set functions, which well matches the permutation invariance property of interference channels and it is robust to imperfect channel state information (CSI). Extensive simulations will show that the proposed IGCNet outperforms existing methods and achieves significant speedup over the classic algorithm for power control, namely, WMMSE.

show abstract

Section: B Existing Approaches' Limitationsmentioning

confidence: 87%

Section: B Existing Approaches' Limitationsmentioning

confidence: 87%

A Graph Neural Network Approach for Scalable Wireless Power Control

Shen

Shi

Zhang

et al. 2019

2019 IEEE Globecom Workshops (GC Wkshps)

View full text Add to dashboard Cite

show abstract

“…The cumulative distribution function (CDF) for P sum in Fig. 9 is obtained over 5000 testing data sets [29]. It is observed that the total transmit power of the IoTDs obtained from the trained DNN is very close to that obtained from the proposed joint optimization while significantly outperforming the disjoint optimizations.…”

Section: Deep Neural Network Evaluationmentioning

confidence: 77%

“…Accordingly, in this section, we present a supervised deep learning method using the DNN to approximate the proposed Algorithm 2, such that by passing the input operating parameters of Algorithm 2 through a trained DNN gives a feasible output for the resource allocation for the C-RAN network with much reduced execution time. Furthermore, training the DNN is fairly convenient as the training samples can easily be obtained by running Algorithm 2 offline [29]. Next, we describe the DNN architecture used in our work, as shown in Fig.…”

Section: Low-complexity Implementation For the Joint Optimization mentioning

confidence: 99%

Computation Offloading for IoT in C-RAN: Optimization and Deep Learning

Pradhan

She

et al. 2020

IEEE Trans. Commun.

View full text Add to dashboard Cite

We consider computation offloading for Internet-of-things (IoT) applications in multiple-inputmultiple-output (MIMO) cloud-radio-access-network (C-RAN). Due to the limited battery life and computational capability in the IoT devices (IoTDs), the computational tasks of the IoTDs are offloaded to a MIMO C-RAN, where a MIMO radio resource head (RRH) is connected to a baseband unit (BBU) through a capacity-limited fronthaul link, facilitated by the spatial filtering and uniform scalar quantization. We formulate a computation offloading optimization problem to minimize the total transmit power of the IoTDs while satisfying the latency requirement of the computational tasks, and find that the problem is non-convex. To obtain a feasible solution, firstly the spatial filtering matrix is locally optimized at the MIMO RRH. Subsequently, we leverage the alternating optimization framework for joint optimization on the residual variables at the BBU, where the baseband combiner is obtained in a closed-form, the resource allocation sub-problem is solved through successive inner convexification, and the number of quantization bits is obtained by a line-search method. As a low-complexity approach, we deploy a supervised deep learning method, which is trained with the solutions to our optimization algorithm. Numerical results validate the effectiveness of the proposed algorithm and the deep learning method.

show abstract

“…To reduce the on-line computational complexity, the idea of "learning to optimize" is proposed for solving variable optimization in [7]. Most recently, a novel framework of using deep learning to find the solution of constrained functional optimization is proposed in [8].…”

Section: State-of-the-artsmentioning

confidence: 99%

Optimizing Wireless Systems Using Unsupervised and Reinforced-Unsupervised Deep Learning

et al. 2020

View full text Add to dashboard Cite

Resource allocation and transceivers in wireless networks are usually designed by solving optimization problems subject to specific constraints, which can be formulated as variable or functional optimization. If the objective and constraint functions of a variable optimization problem can be derived, standard numerical algorithms can be applied for finding the optimal solution, which however incur high computational cost when the dimension of the variable is high. To reduce the online computational complexity, learning the optimal solution as a function of the environment's status by deep neural networks (DNNs) is an effective approach. DNNs can be trained under the supervision of optimal solutions, which however, is not applicable to the scenarios without models or for functional optimization where the optimal solutions are hard to obtain. If the objective and constraint functions are unavailable, reinforcement learning can be applied to find the solution of a functional optimization problem, which is however not tailored to optimization problems in wireless networks. In this article, we introduce unsupervised and reinforced-unsupervised learning frameworks for solving both variable and functional optimization problems without the supervision of the optimal solutions. When the mathematical model of the environment is completely known and the distribution of environment's status is known or unknown, we can invoke unsupervised learning algorithm. When the mathematical model of the environment is incomplete, we introduce reinforcedunsupervised learning algorithms that learn the model by interacting with the environment. Our simulation results confirm the applicability of these learning frameworks by taking a user association problem as an example.

show abstract

Learning to Optimize: Training Deep Neural Networks for Interference Management

Cited by 725 publications

References 16 publications

A Graph Neural Network Approach for Scalable Wireless Power Control

A Graph Neural Network Approach for Scalable Wireless Power Control

Computation Offloading for IoT in C-RAN: Optimization and Deep Learning

Optimizing Wireless Systems Using Unsupervised and Reinforced-Unsupervised Deep Learning

Contact Info

Product

Resources

About