Machine Learning for Large-Scale Optimization in 6G Wireless Networks
Open accessThe DOI matches its Crossref record; 45 of the 45 references it cites were found in a registry or an index.
Abstract
The sixth generation (6G) wireless systems are envisioned to enable the paradigm shift from “connected things” to “connected intelligence”, featured by ultra high density, large-scale, dynamic heterogeneity, diversified functional requirements, and machine learning capabilities, which leads to a growing need for highly efficient intelligent algorithms. The classic optimization-based algorithms usually require highly precise mathematical model of data links and suffer from poor performance with high computational cost in realistic 6G applications. Based on domain knowledge (e.g., optimization models and theoretical tools), machine learning (ML) stands out as a promising and viable methodology for many complex large-scale optimization problems in 6G, due to its superior performance, computational efficiency, scalability, and generalizability. In this paper, we systematically review the most representative “learning to optimize" techniques in diverse domains of 6G wireless networks by identifying the inherent feature of the underlying optimization problem and investigating the specifically designed ML frameworks from the perspective of optimization. In particular, we will cover algorithm unrolling, learning to branch-and-bound, graph neural network for structured optimization, deep reinforcement learning for stochastic optimization, end-to-end learning for semantic optimization, as well as wireless federated learning for distributed optimization, which are capable of addressing challenging large-scale problems arising from a variety of crucial wireless applications. Through the in-depth discussion, we shed light on the excellent performance of ML-based optimization algorithms with respect to the classical methods, and provide insightful guidance to develop advanced ML techniques in 6G networks. Neural network design, theoretical tools of different ML methods, implementation issues, as well as challenges and future research directions are also discussed to support the practical use of the ML model in 6G wireless networks.
Overview
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
The self-check needs the paper’s full text.
Citations
Cites
Human-level control through deep reinforcement learning
Mnih, V., Kavukcuoglu, K., Silver, D., Rusu, A. A., Veness, J., Bellemare, M. G., Graves, A., Riedmiller, M., Fidjeland, A. K., Ostrovski, G., et al. (2015) · Nature · doi:10.1038/nature14236 · Cited by 30,357
Boyd, S., Parikh, N., Chu, E., Peleato, B., & Eckstein, J. (2011) · Foundations and Trends® in Machine Learning · doi:10.1561/2200000016 · Cited by 17,075
The Unreasonable Effectiveness of Deep Features as a Perceptual Metric
Zhang, R., Isola, P., Efros, A. A., Shechtman, E., & Wang, O. (2018) · 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition · doi:10.1109/cvpr.2018.00068 · Cited by 14,623
A Comprehensive Survey on Graph Neural Networks
Wu, Z., Pan, S., Chen, F., Long, G., Zhang, C., & Yu, P. S. (2021) · IEEE Transactions on Neural Networks and Learning Systems · doi:10.1109/tnnls.2020.2978386 · Cited by 11,000 · cited 1×
“Since the graph structure may be changed by the permutation of nodes, a key desideratum for designing GNNs is that the devised GNN should satisfy permutation invariance or permutation equivariance [81], which can precisely capture the internal structure of graph data.”
MethodContinuous control with deep reinforcement learning
Lillicrap, T. P., Hunt, J. J., Pritzel, A., Heess, N., Erez, T., Tassa, Y., Silver, D., & Wierstra, D. (2015) · arXiv · doi:10.48550/arxiv.1509.02971 · Cited by 10,782
Advances and Open Problems in Federated Learning
Kairouz, P., & McMahan, H. B. (2021) · Foundations and Trends in Machine Learning · doi:10.1561/2200000083 · Cited by 7,772 · cited 5×
“, different organizations keep the data of different features corresponding to same batch of customers [159].”
BackgroundMethodInfluentialFederated Machine Learning: Concept and Applications
Yang, Q., Liu, Y., Chen, T., & Tong, Y. (2019) · ACM Transactions on Intelligent Systems and Technology · doi:10.1145/3298981 · Cited by 6,322
Federated learning: Challenges, methods, and future directions
Li, T., Sahu, A. K., Talwalkar, A., & Smith, V. (2020) · IEEE Signal Processing Magazine · doi:10.1109/msp.2020.2975749 · Cited by 6,210 · cited 3×
“FL [157], [158] has been proposed as a prominent distributed ML scheme to effectively solve the model optimization problem in ML over large-scale wireless networks, where each edge device participates in the training process by exchanging model parameters with data kept locally.”
BackgroundMethodShi, Q., Razaviyayn, M., Luo, Z.-Q., & He, C. (2011) · IEEE Transactions on Signal Processing · doi:10.1109/tsp.2011.2147784 · Cited by 2,129 · cited 1×
“The iterative WMMSE algorithm is one of the most representative algorithms for sum-rate maximization problems [61], which is”
MethodA Reduction of Imitation Learning and Structured Prediction to No-Regret Online Learning
Ross, S., Gordon, G. J., & Bagnell, J. A. (2018) · Carnegie Mellon University · doi:10.1184/r1/6550949 · Cited by 1,941
Research works printed but not linked: 245
- 10.1109/COMST.2021.3131332 · DOI · Paper page
- 10.1109/COMST.2021.3130901 · DOI · Paper page
- 10.1109/MCOM.2019.1900271 · DOI · Paper page
- 2022, 6G vision and candidate technologies
- 10.1109/jsac.2023.3280978 · DOI
- 10.1109/JPROC.2021.3061701 · DOI · Paper page
- 10.1109/COMST.2020.3007787 · DOI · Paper page
- 10.1109/MSP.2014.2377273 · DOI · Paper page
- 10.1109/MWC.2015.7143330 · DOI · Paper page
- 10.1109/MNET.121.2100320 · DOI
- Beck, 2022, Semantic communication: An information bottleneck view, arXiv:2204.13366 · arXiv
- 10.1109/twc.2023.3284263 · DOI
- 10.1109/ISIT.2015.7282651 · DOI · Paper page
- 10.1109/TIT.2007.909108 · DOI · Paper page
- 10.1109/TWC.2020.2996144 · DOI
- 10.1109/TWC.2021.3120926 · DOI
- 10.1109/TIT.2008.917634 · DOI
- 10.1109/TSP.2017.2690386 · DOI
- 10.1109/TCOMM.2021.3114791 · DOI · Paper page
- 10.1109/MBITS.2022.3212978 · DOI
- 10.1109/JSAC.2021.3078500 · DOI
- 10.1109/TWC.2021.3100500 · DOI
- 10.1109/COMST.2022.3156207 · DOI
- Amos, 2022, Tutorial on amortized optimization for learning to optimize over continuous domains, arXiv:2202.00665 · arXiv
- 10.1109/ACCESS.2022.3218802 · DOI · Paper page
- 10.1109/TWC.2022.3219840 · DOI · Paper page
- 10.1109/TCOMM.2019.2924010 · DOI · Paper page
- 10.1109/TSP.2020.2976585 · DOI · Paper page
- 10.1109/TCOMM.2020.3027027 · DOI · Paper page
- 10.1109/TWC.2022.3181219 · DOI
- Khan, 2018, Scalable centralized deep multi-agent reinforcement learning via policy gradients, arXiv:1805.08776 · arXiv
- 10.1109/MWC.006.2100543 · DOI · Paper page
- 10.1109/TSP.2022.3143372 · DOI
- 10.1109/TWC.2020.3030882 · DOI
- 10.1109/TCOMM.2021.3135540 · DOI
- 10.1109/MCOM.001.2001239 · DOI · Paper page
- 10.1109/MCOM.101.2100138 · DOI
- 10.1109/LCOMM.2019.2898944 · DOI · Paper page
- Haykin, 2004, A comprehensive foundation, Neural Netw., 2, 41
- 10.1109/TVT.2019.2897134 · DOI · Paper page
- 10.1109/MSP.2020.3016905 · DOI · Paper page
- 10.1109/OJCOMS.2022.3210648 · DOI
- 10.1109/COMST.2019.2904897 · DOI · Paper page
- 10.1109/COMST.2019.2926625 · DOI · Paper page
- 10.1109/JSAC.2019.2933969 · DOI · Paper page
- 10.1109/MWC.001.2200023 · DOI
- 10.1109/COMST.2022.3223224 · DOI · Paper page
- 10.1109/COMST.2019.2916583 · DOI · Paper page
- 10.1109/COMST.2021.3089688 · DOI · Paper page
- 10.1109/COMST.2022.3199544 · DOI · Paper page
- 10.1109/COMST.2020.2988367 · DOI · Paper page
- 10.1109/TSP.2021.3079807 · DOI
- 10.1109/TWC.2021.3123620 · DOI
- 10.1109/JSAC.2021.3078502 · DOI · Paper page
- 10.1109/TWC.2020.3040983 · DOI · Paper page
- 10.1109/TSP.2022.3163626 · DOI
- 10.1109/TITS.2019.2963722 · DOI · Paper page
- 10.1109/LCOMM.2020.2990487 · DOI
- 10.1109/APCC.2018.8633520 · DOI
- 10.1109/ICC.2017.7997286 · DOI · Paper page
- 10.1109/ACCESS.2019.2959348 · DOI
- 10.1109/JIOT.2021.3058236 · DOI
- 10.1109/JSAC.2020.3036968 · DOI · Paper page
- 10.1109/JSAC.2022.3221990 · DOI · Paper page
- 10.1109/JSAC.2022.3221977 · DOI · Paper page
- 10.1109/JSAC.2021.3087240 · DOI · Paper page
- 10.1109/jsac.2022.3221991 · DOI · Paper page
- 10.1109/LCOMM.2020.3019312 · DOI
- 10.1109/TWC.2021.3128392 · DOI · Paper page
- 10.1109/INFOCOM41043.2020.9155494 · DOI · Paper page
- 10.1109/ICC45855.2022.9838703 · DOI
- Gregor, Learning fast approximations of sparse coding, 399
- 10.1109/TSP.2017.2708040 · DOI · Paper page
- Chen, 2021, Learning to optimize: A primer and a benchmark, arXiv:2103.12828 · arXiv
- 10.1109/TPAMI.2018.2883941 · DOI · Paper page
- 10.1109/ICASSP.2019.8682542 · DOI
- 10.1109/TCOMM.2021.3114682 · DOI
- 10.1109/TWC.2022.3142737 · DOI
- 10.1109/JIOT.2021.3097133 · DOI
- 10.1109/TWC.2021.3072382 · DOI
- 10.1109/TWC.2022.3166964 · DOI
- 10.1561/2000000093 · DOI · Paper page
- 10.1109/18.771234 · DOI
- 10.1109/TCI.2020.2964202 · DOI · Paper page
- 10.2307/1910129 · DOI · Paper page
- Huang, 2021, Branch and bound in mixed integer linear programming problems: A survey of techniques and trends, arXiv:2111.06257 · arXiv
- 10.1287/ijoc.2016.0723 · DOI · Paper page
- 10.1016/j.disopt.2016.01.005 · DOI · Paper page
- Gasse, Exact combinatorial optimization with graph convolutional neural networks, 15554
- Nair, 2020, Solving mixed integer programs using neural networks, arXiv:2012.13349 · arXiv
- He, Learning to search in branch and bound algorithms, 27, 3293
- 10.3390/ai2020010 · DOI
- 10.1109/49.793310 · DOI · Paper page
- 10.1109/TWC.2013.040413.120676 · DOI · Paper page
- 10.1109/TVT.2019.2953724 · DOI
- 10.2307/353415 · DOI · Paper page
- 10.1007/s10851-007-0008-7 · doi:10.1007/s10851-007-0008-7
- 10.1109/IJCNN.2010.5596450 · DOI
- 10.1109/LCOMM.2022.3159742 · DOI
- 10.1109/TSP.2017.2699640 · DOI
- 10.1109/mnet.123.2100773 · DOI
- 10.1109/GCWkshps45667.2019.9024538 · DOI
- 10.1109/TWC.2022.3193138 · DOI
- 10.1109/ICASSP39728.2021.9414098 · DOI
- 10.1109/ICASSP43922.2022.9746926 · doi:10.1109/icassp43922.2022.9746926
- 10.1109/TSP.2023.3255547 · DOI
- 10.1109/TMC.2021.3125793 · DOI
- 10.1109/TWC.2021.3100133 · DOI
- 10.1109/TMC.2020.3020582 · DOI
- 10.1109/TMC.2021.3079117 · DOI
- 10.1109/GLOBECOM46510.2021.9685457 · DOI
- Garg, Generalization and representational limits of graph neural networks, 3419
- Alon, On the bottleneck of graph neural networks and its practical implications, 1
- Bertsekas, 1995, Dynamic Programming and Optimal Control
- 10.1109/TITS.2019.2906260 · DOI
- 10.1145/3219819.3220096 · DOI · Paper page
- 10.1109/WCNC49053.2021.9417262 · DOI
- 10.1109/ITSC.2018.8569568 · DOI · Paper page
- 10.1109/ITSC.2017.8317839 · DOI · Paper page
- 10.1016/j.trc.2019.08.011 · DOI · Paper page
- 10.1016/j.energy.2021.122523 · DOI
- 10.1109/VPPC.2018.8605023 · DOI
- 10.1109/ACCESS.2020.2991057 · DOI
- 10.1109/JIOT.2019.2928837 · DOI
- 10.1109/LWC.2021.3052079 · DOI
- 10.1145/3132847.3133045 · DOI
- 10.1109/MICRO50266.2020.00090 · DOI · Paper page
- 10.1145/3288599.3288634 · DOI
- 10.1109/TGCN.2020.3045075 · DOI
- 10.1109/CSCloud-EdgeCom52276.2021.00044 · DOI
- 10.1109/LCOMM.2020.3001227 · DOI
- 10.1109/LCOMM.2021.3053612 · DOI
- 10.1109/TWC.2020.2993071 · DOI · Paper page
- 10.1109/TPDS.2021.3132422 · DOI
- 10.1109/TNSE.2021.3068340 · DOI · Paper page
- 10.1109/TITS.2021.3128209 · DOI
- 10.1109/TII.2022.3164395 · DOI
- 10.1109/MWC.001.00254 · DOI
- 10.1109/JSTSP.2021.3135751 · DOI
- 10.1109/TWC.2020.3001736 · DOI · Paper page
- 10.1109/TITS.2021.3054511 · DOI
- 10.1016/j.jet.2004.03.008 · DOI · Paper page
- Littman, A generalized reinforcement-learning model: Convergence and applications, 96, 310
- Silver, Deterministic policy gradient algorithms, 387
- 10.1109/JSYST.2019.2933536 · DOI
- 10.1145/3466752.3480129 · DOI
- 10.1145/584091.584093 · DOI
- 10.1002/j.1538-7305.1948.tb01338.x · DOI · Paper page
- 10.1109/MWC.101.2100269 · DOI · Paper page
- 10.1017/9781108779302 · doi:10.1017/9781108779302
- Chechik, 2005, Information bottleneck for Gaussian variables, J. Mach. Learn. Res., 6, 165
- 10.1109/JSAIT.2020.2991561 · DOI · Paper page
- Alemi, Deep variational information bottleneck, 1
- Kingma, Auto-encoding variational Bayes, 1
- 10.1109/TCCN.2022.3173056 · DOI
- 10.1109/TCCN.2019.2919300 · DOI · Paper page
- 10.1109/JSAIT.2020.2987203 · DOI · Paper page
- 10.1109/JSAC.2022.3221968 · DOI · Paper page
- 10.1109/ICCWorkshops53468.2022.9814492 · DOI
- 10.1109/JSAC.2022.3191326 · DOI · Paper page
- 10.1109/TCSVT.2021.3082521 · DOI · Paper page
- 10.1109/JSAC.2022.3221999 · DOI · Paper page
- 10.1109/twc.2023.3240969 · DOI · Paper page
- 10.1109/JSAC.2022.3180802 · DOI · Paper page
- 10.1109/MWC.008.2200180 · DOI
- 10.1109/MCOM.001.2000604 · DOI · Paper page
- 10.1109/COMST.2021.3058573 · DOI · Paper page
- 10.1109/JIOT.2022.3168066 · DOI
- 10.1109/JSAC.2022.3180799 · DOI
- 10.1109/TNSE.2022.3178970 · DOI · Paper page
- 10.1109/TCOMM.2019.2944169 · DOI · Paper page
- 10.1109/TWC.2021.3062708 · DOI · Paper page
- 10.1109/TWC.2019.2946245 · DOI · Paper page
- Wang, 2022, Over-the-air computation: Foundations, technologies, and applications, arXiv:2210.10524 · arXiv
- 10.1109/TSP.2022.3171065 · doi:10.1109/tsp.2022.3171065
- 10.1109/TWC.2022.3227636 · DOI
- 10.1109/TII.2021.3064351 · DOI · Paper page
- 10.1109/JIOT.2020.3026589 · DOI · Paper page
- 10.1109/JIOT.2022.3174469 · DOI
- 10.1109/TCOMM.2020.3004524 · DOI
- 10.1109/TVT.2021.3131534 · DOI
- 10.1109/ACCESS.2021.3095915 · DOI
- 10.1109/VTC2021-Spring51267.2021.9448736 · DOI
- 10.1109/ICC42927.2021.9500603 · DOI
- 10.1109/GLOBECOM42002.2020.9348225 · DOI
- 10.1109/LNET.2022.3173971 · DOI
- Liu, A communication efficient vertical federated learning framework, 118
- 10.1137/1.9781611972726.18 · doi:10.1137/1.9781611972726.18
- 10.1109/MCOM.001.2001016 · DOI
- 10.1109/msp.2023.3262366 · doi:10.1109/msp.2023.3262366
- 10.1109/TVT.2019.2951501 · DOI · Paper page
- 10.1109/MNET.2019.1900029 · DOI
- 10.1109/LCOMM.2020.3035326 · DOI
- 10.1109/MCOM.001.2000373 · DOI · Paper page
- 10.1109/TSP.2019.2908906 · DOI · Paper page
- 10.1109/TWC.2020.2992786 · DOI
- 10.1109/JPROC.2021.3053601 · DOI · Paper page
- 10.1109/ICC.2019.8761681 · DOI
- 10.1109/JSAC.2020.3036947 · DOI
- 10.1109/tnnls.2023.3283523 · DOI · Paper page
200 of 245 shown
Cited by
Modern computing: Vision and challenges
Gill, S. S., Wu, H., Patros, P., Ottaviani, C., Arora, P., Pujol, V. C., Haunschild, D., Parlikad, A. K., Cetinkaya, O., Lutfiyya, H., et al. (2024) · Telematics and Informatics Reports · doi:10.1016/j.teler.2024.100116 · Cited by 197
“Both B5G and 6G networks aim to enable the smooth and complete integration of many industries, including the IoT, aerial networks (also known as drones), satellite accessibility, and submerged connectivity [98].”
BackgroundInfluentialYang, C., Kwong, C. F., Chieng, D., Kar, P., Yau, K.-L. A., & Chen, Y. (2025) · IEEE Communications Surveys & Tutorials · doi:10.1109/comst.2024.3440033 · Cited by 33
Metadata
- Authors
- Shi, Y., Lian, L., Shi, Y., Wang, Z., Zhou, Y., Fu, L., Bai, L., Zhang, J., & Zhang, W.
- Venue
- IEEE Communications Surveys & Tutorials
- Published
- 2023
- Volume
- 25
- Issue
- 4
- Pages
- 2088-2132
- ISSN
- 1553-877X (ISSN-L)
- 2373-745X
- OpenAlex
- W4385453555
- Cited by
- 205
- References
- 45
- FWCI
- 15.03
- Type
- journal article
- Language
- en
- Open access
- green
- Licence
- Funders
- National Nature Science Foundation of China (62101331, 62271318)
- Natural Science Foundation of Shanghai (21ZR1442700, 23ZR1442800)
- Shanghai Rising-Star Program (22QA1406100)
- National Natural Science Foundation of China (U20A20159, 62001294)
- Research Impact Fund (RIF) from Hong Kong RGC (R5009-21)
- Special Project for Research and Development in Key areas of Guangdong Province (2020B0101110003)
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
How to cite
Cite CoChat for the checks on this page, and the paper for its own findings.
- This page
- CoChat. (2026). Citation check: Machine Learning for Large-Scale Optimization in 6G Wireless Networks. https://app.cochat.ai/papers/10.1109/comst.2023.3300664
- The paper
- Shi, Y., Lian, L., Shi, Y., Wang, Z., Zhou, Y., Fu, L., Bai, L., Zhang, J., & Zhang, W. (2023). Machine Learning for Large-Scale Optimization in 6G Wireless Networks. IEEE Communications Surveys & Tutorials. https://doi.org/10.1109/comst.2023.3300664
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.