Approximate strategies for data recovery and transmission in unstable communication networks

다국어 초록 (Multilingual Abstract)

As Artificial Intelligence (AI) is expected to be integrated into sixth-generation (6G) core technologies, the management and transmission of a massive volume of data are becoming increasingly crucial. These changes are characterized by a significant increase in data, encompassing Machine Learning (ML) models distributed among numerous devices, underscoring the importance of data in interconnected, data-driven environments. Therefore, guaranteeing data reliability and reducing the impact of network instability on data delivery becomes crucial. In environments with changing network conditions, the complete data delivery and recovery have been an ultimate goal. However, a manageable amount of data loss is acceptable for error-tolerant applications or AI model training without significant effects on user experience. This indicates that perfect data delivery and recovery methods for such data could actually lead to inefficient use of network resources. Therefore, a new shift is necessary towards transmission methods that enhance network resource efficiency and ensure uninterrupted service, aligned with the fundamental characteristics of data. To design a resilient and efficient network, we must address several critical issues, which are data loss incurred by network instability, network congestion from multiple users, and ensuring network efficiency. Addressing these issues requires novel data transmission strategies in communication networks. To overcome data loss, we propose a parameter transferred irreducible Long Short-Term Memory (LSTM) algorithm to accurately impute the missing values. For a network congestion problem, we propose a method that leverages Network Coding (NC) by recovering lost packets through other packets without the need for retransmissions, thereby preventing network congestion and enabling the establishment of a reliable network. For network efficiency, we extend NC to include low-rank approximation, addressing the all-or-nothing challenge that often requires extra packet transmissions in traditional NC scenarios. This modification optimizes not only the transmission process for improved efficiency but also reduces the computational complexity of the NC decoding process. We use this algorithm to integrate with federated learning, proposing a solution that mitigates learning delays and performance degradation due to packet loss in both directions of communication during the local model training process. To validate the enhanced performance of our proposed algorithms, we assess them across diverse real-world and synthetic datasets. It is demonstrated that our algorithms outperform the current advanced methods in addressing key challenges such as missing data, network congestion, network efficiency. Additionally, these proposed algorithms show effective results against the specific performance metrics of each application, a critical factor in developing resilient and efficient networks.

국문 초록 (Abstract)

인공지능이 6세대 통신의 핵심 기술에 통합될 것으로 예상됨에 따라, 대량의 데이터 관리와 전송의 역할이 커지고 있다. 기계 학습 모델을 포함해 다양한 데이터의 양이 급격히 증가하고 있으며, 이는 데이터의 중요성을 더욱 강조하고 있다. 이에 따라, 불안정한 네트워크가 데이터 전송에 끼치는 영향을 줄이고 데이터 신뢰성을 유지하는 것이 필요하다. 기존 연구에서는 불안정한 네트워크 환경에서 완벽한 데이터 전송 및 복구를 궁극적 목표로 삼았다. 그러나 오류에 강인한 애플리케이션이나 인공지능 모델 학습의 경우, 사용자 경험에 큰 영향을 주지 않는 범위 내에서 일정량의 데이터 손실이 허용되기 때문에 완벽한 데이터 전송 및 복구 방식이 오히려 네트워크 자원을 비효율적으로 사용하게 만들 수 있음을 시사한다. 따라서, 완벽한 통신 방법을 추구하기보다는 네트워크 리소스의 효율성을 향상시키고 중단 없는 서비스를 보장함으로써 데이터의 기본적인 특성에 맞는 신뢰할 수 있는 새로운 전송 방식으로 전환하는 것이 필요하다. 탄력적이고 효율적인 네트워크를 설계하기 위해서는 네트워크 불안정성으로 인한 데이터 손실, 다수 사용자로 인한 네트워크 혼잡, 네트워크 효율성을 보장하는 여러 중요한 문제를 해결해야 한다. 데이터 손실을 극복하기 위해, 본 학위 논문에서는 누락된 값을 높은 정확도로 대체할 수 있는 파라미터 전송 LSTM 알고리즘을 제안한다. 네트워크 혼잡 문제에 대해서는, 다른 패킷을 통해 분실된 패킷을 복구함으로써 재전송의 필요성 없이 네트워크 혼잡을 방지하고 신뢰할 수 있는 네트워크를 구축할 수 있는 네트워크 코딩을 활용하는 방법을 제안한다. 마지막으로 네트워크 효율성을 향상시키기 위해, 본 학위 논문에서는 전통적인 네트워크 코딩 시나리오에서 발생하는 ``All-or-Nothing'' 문제를 해결하기 위해 네트워크 코딩 기법에 저순위 근사법(low-rank approximation)을 결합한 새로운 데이터 전송 방법을 제안한다. 본 학위 논문에서는 제안한 알고리즘의 성능 향상을 검증하기 위해 다양한 실제 및 합성 데이터셋을 통해 성능을 평가하였다. 다양한 실험을 통하여 제안한 알고리즘이 데이터 전가, 네트워크 혼잡, 네트워크 효율성 측면에서 최신 기술들보다 우수한 성능을 보임을 확인하였다. 또한, 각 애플리케이션의 특정 성능 지표에 대해 효과적인 성능을 보임을 실험적으로 입증했으며, 이는 제안된 알고리즘이 불안정한 네트워크 환경에서 탄력적이고 효율적인 네트워크를 개발하는 데 유용한 기술임을 시사한다.

목차 (Table of Contents)

• 1. Introduction 1
• I. Motivation and Goal of Dissertation 1
• II. Contributions 2
• 2. Parameter Transferred Irreducible LSTM for Data Imputation 7
• I. Introduction 7
• II. Related Works 11
• A. Types of Missing Traffic Data in Traffic Environment 11
• B. Traffic Data Imputation Methods . 12
• III. System Setup and Problem Formulation 15
• IV. Proposed Designed Architecture 16
• A. Data Selection 17
• B. Extrapolation by LSTM Factorization with Parameter Transfer. 18
• C. Spatial Interpolation 22
• V. Experiments. 24
• A. Experiment Setup 24
• B. Performance Evaluation using Real-World Traffic Dataset 26
• C. Performance Evaluation using Synthetic Dataset 32
• VI. Conclusions 36
• 3. Data Dissemination Framework using Low-rank Approximation in Edge Networks 42
• I. Introduction 42
• II. Related Works 46
• A. Reliable Data Dissemination Approach 46
• B. Low-ComplexityNetworkCoding 47
• III. System Setup andProblemFormulation 50
• IV. Data Dissemination Framework 52
• A. Encoding with Matrix Decomposition 52
• B. Decoding with Low-rank Approximation 53
• C. Complexity Analysis 54
• D. Procedures of Proposed Equal Data Dissemination 55
• V. Analysis of Decoding Accuracy 55
• A. Comparison of Decoding Accuracy 55
• B. Upper Bound of Decoding Accuracy 59
• VI. Simulation Results 63
• A. Simulation Setup 63
• B. Performance Evaluation 67
• C. The Impact of Data Correlation on Decoding Accuracy 68
• D. Performance Comparison with Packet Loss 69
• E. Complexity Comparison with Network Coding Algorithms 71
• VII. Conclusions 72
• VIII. Appendix 72
• A. Validation of Upper Bound 72
• 4. Resilient and Efficient Approximate Packet Recovery for Federated Learning 82
• I. Introduction. 82
• II. Related Works 87
• A. Low-rank Approximation in FL 87
• B. FL with Network Instability 87
• C. Systematic Network Coding 89
• III. System Setup and Problem Formulation 95
• IV. Low-Rank Approximation Transmission Procedure for FL 97
• A. Low-rank Model Parameter Approximation 98
• B. Robust SysNC for Low-Rank Model Parameters . 101
• C. OverallProcedure 105
• V. Theoretical Analysis 106
• A. Impact of Packet Loss on Model Approximation 106
• B. Convergence Analysis 108
• VI. Experiments 111
• A. Experiment Setup 112
• B. Performance Comparison in Packet Erasure Network 114
• C. Evaluation of Hybrid FL in Packet Erasure Network 116
• D. Analysis of S-Dynamic Singular Values 118
• VII. Conclusions 122
• VIII. Appendix 123
• A. Proof of Theorem 5 123
• B. KeyLemmas 125
• 5. Implementation of Network Coding Algorithm for Single Queue with Single Memory Devices 132
• I. Introduction 132
• II. SystemSetup 133
• A. Overview of Network Coding Operations 133
• B. Network Modeland Assumption 135
• III. ProposedAlgorithms 136
• A. Proposed Encoding and Decoding Processes 136
• IV. Implementation Results 137
• V. Conclusion 138
• 6. Traffic Data Classification using Machine Learning Algorithms in SDN Networks 142
• I. Introduction 142
• II. System Setup 144
• III. Experiment Results 147
• A. Experiment Setup 147
• B. Evaluation 147
• IV. Conclusion 149
• 7. Conclusions 150
• References 153
• Abstract in Korean 180
• Acknowledgement in Korean 182

참고문헌 (Reference)

1. HTTP/3, RFC 9114, M. Bishop, Internet Engineering Task Force doi10.17487/RFC9114Online Available https//www rfceditor. org/rfc/rfc9114, , 2022

2. 143 Korea highway system, Korea Expressway Corporation, from http://data. ex. co. kr/portal/fdwn/view?type=VDS&num =38&requestfromd%CC%84atasetAccessed, , 2018

3. Topics in matrix analysis, C. R. Johnson, R. A. Horn, R. A. Horn and, Cambridge University Press, , 1994

4. Federated mutual learning,, J. Zhang, X. Jia et al., T. Shen, arXiv preprint arXiv:2006.16765, , 2020

5. Application-awareness in SDN, in, G. Bellala, M. Arndt and, Z. A. Qazi, T. Jin, J. Lee, G. Noubir, Proceedings of the Association for Computing Machineryz SIGCOMM, vol. 43, pp. 487–488, , 2013

6. IPv6 Segment Routing Header (SRH), C. Filsfils, D. Dukes, S. Matsushima and, J. Leddy, S. Previdi, E. Vyncke, RFCOnline Available https//www rfc-editor. org/rfc/rfc8754. html, , 2020

7. Network intelligence technologies, H. Kim, M. Shin, B. Ahn et al., ETRI Insight, , 2018

8. Sliding network coding for URLLC,, J. Choi, vol. 21, no. 6, pp. 4424–4433, , 2021

9. Advanced message queuing protocol,, S. Vinoski, IEEE Internet Computing, vol. 10, no. 6, pp. 87–89, , 2006

10. Edge computingVision and challenges,, Q. Zhang, W. Shi, Y. Li and, L. Xu, J. Cao, vol. 3, no. 5, pp. 637– 646, , 2016

11. Modelling using polynomial regression,, E. Ostertagov´a, Procedia Engineering, vol. 48, pp. 500–506, , 2012

12. AdamA method for stochastic optimization, J. Ba, D. P. Kingma and, arXiv preprint arXiv:1412.6980, , 2014

13. Federated learning with packet losses, in, A. Rodio, G. Neglia, F. Busacca et al., 2023 26th International Symposium on Wireless Personal Multimedia Communications (WPMC), IEEE, 2023, pp. 1–6, , 2023

14. Aspects of multivariate statistical theory, R. J. Muirhead, John Wiley & Sons vol. 197, , 2009

15. MQTT Essentials-A lightweight IoT protocol, G. C. Hillar, Packt Publishing Ltd, , 2017

16. On the convergence of fedavg on non-iid data, Z. Zhang, S. Wang and, X. Li, W. Yang, K. Huang, arXiv preprint arXiv:1907.02189, , 2019

17. Pca learning for sparse high-dimensional data,, M. Rattray, D. C. Hoyle and, vol. 62, no. 1, p. 117, , 2003

18. Resource scheduling in edge computingA survey,, W. Shi, S. Hu, G. Li and, C. Li, Q. Luo, IEEE Communications Surveys & Tutorials, vol. 23, no. 4, pp. 2131–2165, , 2021

19. A first look at HTTP/3 adoption and performance,, I. Drago, M. Trevisan, G. Perna, D. Giordano and, Computer Communications, vol. 187, pp. 115–124, , 2022

20. XORs in the airPractical wireless network coding, W. Hu, H. Rahul, S. Katti, M. Medard and, J. Crowcrof, D. Katabi, IEEE Transactions on networking, vol. 16, no. 3, pp. 497–510, , 2008

21. Xors in the airPractical wireless network coding, D. Katabi, W. Hu, S. Katti, J. Crowcroft, H. Rahul, M. M´edard and, IEEE/ACM Transactions on Networking, vol. 16, no. 3, pp. 497–510, , 2008

22. Federated learning over wireless fading channels,, M. M. Amiri and, D. G¨und¨uz, IEEE Transactions on Wireless Communications, vol. 19, no. 5, pp. 3546–3557, , 2020

23. Flow-based intrusion detection system for SDN, in, I. H. Elhajj, A. Kayssi and, N. Adalian, A. Chehab, G. A. Ajaeiya, IEEE Symposium on Computers and Communications (ISCC), pp. 787–793, , 2017

24. Low rank communication for federated learning, in, B. Jin, X. Wang and, H. Zhou, J. Cheng, Database Systems for Advanced Applications 2020 International Workshops, Springer, pp. 1–16, , 2020

25. Why random pruning is all we need to start sparse, A. H. Gadhikar, S. Mukherjee and, R. Burkholz, International Conference on Machine Learning (ICML), PMLR, 2023, , 2023

26. Automated Synoptic Observing System (ASOS) Dataset, K. M. AKMA, Last accessedOnline Available https://data. kma. go. kr/data/grnd/selectAsosRltmList. do?pgmNo=36, , 2021

27. Admission control with online algorithms in SDN, in, S. Chouvardas, L. Maggi, J. Leguay, S. Paris and, M. Draief, IEEE Network Operations and Management Symposium (NOMS), pp. 718–721, , 2016

28. P8P4 with predictable packet processing performance, M. Jarschel, W. Kellerer, R. Pries and, M. He, H. Harkous, IEEE Transactions on Network and Service Management, vol. 18, no. 3, pp. 2846–2859, , 2020

29. Learning to forgetContinual prediction with LSTM, in, F. Cummins, F. A. Gers, J. Schmidhuber and, IET International Conference on Artificial Neural Networks, pp. 850–855, , 1999

30. MobileNetV2Inverted Residuals and Linear Bottlenecks, M. Sandler, A. Howard, M. Zhu, A. Zhmoginov and, L. C. Chen, Proceedings of the IEEE conference on Computer Vision and Pattern Recognition, pp. 4510–4520, , 2018

31. Eigenvalues and condition numbers of random matrices,, E. Alan, vol. 9, no. 4, pp. 543– 560, , 1988

32. Missing traffic dataComparison of imputation methods,, Y. Li, L. Li, Z. Li and, IET Intelligent Transport Systems, vol. 8, no. 1, pp. 51–57, , 2014

33. An eigenanalysis of data centering in machine learning, P. Honeine, arXiv preprint arXiv:1407.2904, , 2014

34. Learning multiple layers of features from tiny images,, A. Krizhevsky, G. Hinton et al., Technical report, , 2009

35. Systematic network coding for time-division duplexing,, D. E. Lucani, M. M´edard and, M. Stojanovic, in 2010 IEEE International Symposium on Information Theory (ISIT), IEEE pp. 2403–2407, , 2010

36. BRITSBidirectional recurrent imputation for time series, W. Cao et al., in Conference on Neural Information Processing Systems, pp. 6775–6785, , 2018

37. MICE: Multivariate imputation by chained equations in r, K. Groothuis-Oudshoorn, S. Van Buuren and, vol. 45, pp. 1–67, , 2011

38. TensorflowA system for large-scale machine learning, in, M. Abadi et al., 12th USENIX Symposium on Operating Systems Design and Implementation, pp. 265–283, , 2016

39. Convergence of federated learning over a noisy downlink,, D. G¨und¨uz, M. M. Amiri, H. V. Poor, S. R. Kulkarni and, IEEE Transactions on Wireless Communications, vol. 21, no. 3, pp. 1422–1437, , 2021

40. SplitfedWhen federated learning meets split learning, in, C. Thapa, P. C. M. Arachchige, S. Camtepe and, L. Sun, Proceedings of the AAAI Conference on Artificial Intelligence, vol. 36, 2022, pp. 8485– 8493, , 2022

41. Revisiting weight initialization of deep neural networks,, A. Temperoni and, M. Skorski, M. Theobald, Asian Conference on Machine Learning (ACML), PMLR, pp. 1192–1207, , 2021

42. Sparse communication for distributed gradient descent, in, K. Heafield, A. F. Aji and, Proceedings of the 2017 Conference on Empirical Methods in Natural Language Processing (EMNLP), pp. 440–445, , 2017

43. Cautious view on network coding—from theory to practice,, T. Larsen, J. Heide, F. H. P. Fitzek and, M. V. Pedersen, Journal of Communications and Networks, vol. 10, no. 4, pp. 403–411, , 2008

44. An intelligent SDN framework for 5G heterogeneous networks,, S. Sun, K. Lu, B. Rong and, L. Gong, IEEE Communications Magazine, vol. 53, no. 11, pp. 142–147, , 2015

45. Distribution of eigenvalues for some sets of random matrices,, V. A. Marchenko and, L. A. Pastur, Matematicheskii Sbornik, vol. 114, no. 4, pp. 507–536, , 1967

46. Z-TORCHAn automated NFV orchestration and monitoring solution, X. Costa Perez, F. Z. Yousaf and, V. Sciancalepore, IEEE Transactions on Network and Service Management, vol. 15, no. 4, pp. 1292– 1306, , 2018

47. FSWFulcrum sliding window coding for low-latency communication, G. T. Nguyen, V. Nguyen, E. Tasdemir, M. Reisslein, F. H. Fitzek and, IEEE Access, vol. 10, pp. 54 276–54 290, , 2022

48. Federated learning: Challenges, methods, and future directions, A. Talwalkar and, T. Li, V. Smith, A. K. Sahu, IEEE Signal Processing Magazine, vol. 37, no. 3, pp. 50–60, , 2020

49. On the performance bounds of practical wireless network coding,, J. C. Lui and, D.-M. Chiu, J. Le, IEEE Transactions on Mobile Computing, vol. 9, no. 8, pp. 1134–1146, , 2010

50. Outdoor MIMO wireless channelsModels and performance prediction, A. J. Paulraj, D. A. Gore and, H. Bolcskei, D. Gesbert, IEEE Transactions on Communications, vol. 50, no. 12, pp. 1926–1934, , 2002

51. A fast network coding scheme for mobile wireless sensor networks, C. Han, X. Han, Y. Yang and, International Journal of Distributed Sensor Networks, vol. 13, no. 2, p. 1 550 147 717 693 241, , 2017

52. A survey of packet loss recovery techniques for streaming audio,, C. Perkins, O. Hodson and, V. Hardman, IEEE Network, vol. 12, no. 5, pp. 40–48, , 1998

53. Decentralized federated learning with unreliable communications,, L. Liang and, G. Y. Li, H. Ye, IEEE Journal of Selected Topics in Signal Processing, vol. 16, no. 3, pp. 487–500, , 2022

54. Freeway performance measurement systemMining loop detector data,, C. Chen et al., vol. 1748, no. 1, pp. 96–102, , 2001

55. Temporal belief memoryImputing missing data during RNN training., Y.-J. Kim and, M. Chi, in Proceedings of the 27th International Joint Conference on Artificial Intelligence, , 2018

56. Proactive failure recovery for NFV in distributed edge computing,, H. Huang and, S. Guo, IEEE Communications Magazine, vol. 57, no. 5, pp. 131–137, , 2019

57. Model pruning enables efficient federated learning on edge devices, Y. Jiang, V. Valls, et al., S. Wang, IEEE Transactions on Neural Networks and Learning Systems, , 2022

58. Federated learningStrategies for improving communication efficiency, F. X. Yu, P. Richt´arik, D. Bacon, J. Konevcn`y, H. B. McMahan, A. T. Suresh and, arXiv preprint arXiv:1610.05492, , 2016

59. FetchsgdCommunicationefficient federated learning with sketching in, E. Ullah et al., A. Panda, D. Rothchild, International Conference on Machine Learning (ICML), PMLR, pp. 8253–8265, , 2020

60. A multi-hop data dissemination algorithm for vehicular communication,, H. Kim, O. Urmonov and, Computers, vol. 9, no. 2, p. 25, , 2020

61. Federated learning for 6GApplications, challenges, and opportunities,, S. Cui, M. Chen, K.-K. Wong, H. V. Poor and, Z. Yang, Engineering, vol. 8, pp. 33–41, , 2022

62. On the importance of initialization and momentum in deep learning, in, G. Dahl and, I. Sutskever, J. Martens, G. Hinton, International Conference on Machine Learning (ICML), PMLR, pp. 1139–1147, , 2013

63. QueenEstimating packet loss rate between arbitrary internet hosts, in, J. Li and, C. Huang, K. W. Ross, Y. A. Wang, Passive and Active Network Measurement10th International Conference, PAM 2009, Seoul, Korea, Proceedings 10, Springer, pp. 57– 66, , 2009

64. An efficient realization of deep learning for traffic data imputation,, F.-Y. Wang, Y.-L. Liu and, Y. Lv, Y. Duan, Transportation research part Cemerging technologies, vol. 72, pp. 168–181, , 2016

65. Communication-efficient federated learning via knowledge distillation,, X. Xie, Y. Huang and, L. Lyu, F. Wu, C. Wu, vol. 13, no. 1, p. 2032, , 2022

66. New algorithm of multi-strategy channel allocation for edge computing,, H. Zhu, C. Chen and, T. Zhang, M. Piao, D. Zhang, AEU-International Journal of Electronics and Communications, vol. 126, p. 153 372, , 2020

67. Performance optimization of federated learning over wireless networks,, W. Saad, H. V. Poor and, Z. Yang, C. Yin, S. Cui, M. Chen, in 2019 IEEE global communications conference (GLOBECOM), IEEE, pp. 1–6, , 2019

68. Online outlier detection in sensor data using non-parametric models, in, S. Subramaniam et al., Proceedings of the 32nd International Conference on Very Large Data Bases, pp. 187–198, , 2006

69. FedALAAdaptive local aggregation for personalized federated learning, in, J. Zhang, H. Wang et al., Y. Hua, Proceedings of the AAAI Conference on Artificial Intelligence, vol. 37, 2023, pp. 11 237–11 244, , 2023

70. Graph convolutional networks for traffic forecasting with missing values,, J. Zuo et al., Data Mining and Knowledge Discovery, vol. 37, no. 2, pp. 913–947, 2023, , 2023

71. A low-complexity coded transmission scheme over finite-buffer relay links,, S. Zhang, J. Wang, X. Ji, H. Wu and, Y. Li, Z. Bao, IEEE Transactions on Communications, vol. 66, no. 7, pp. 2873–2887, , 2018

72. An applicable repeated transmission for low latency and reliable services,, Y. Lee and, H. Lee, E. Kim, IEEE Transactions on Vehicular Technology, vol. 69, no. 8, pp. 8468–8482, , 2020

73. Compressed network codingOvercome all-or-nothing problem in finite fields,, M. Kwon, H. Park and, P. Frossard, in 2014 IEEE Wireless Communications and Networking Conference (WCNC), IEEE, pp. 2851–2856, , 2014

74. Limiting form of the sample covariance eigenspectrum in PCA and kernel PCA, D. Hoyle and, M. Rattray, Advances in Neural Information Processing Systems, vol. 16, , 2003

75. Pruning randomly initialized neural networks with iterative randomization,, T. Inoue, S. Yamaguchi, D. Chijiwa, K. Umakoshi and, Y. Ida, Advances in neural information processing systems, vol. 34, pp. 4503– 4513, , 2021

76. Spatio-temporal wireless traffic prediction with recurrent neural network,, P. Zhang and, Z. Feng, C. Qiu, Y. Zhang, S. Cui, vol. 7, no. 4, pp. 554–557, , 2018

77. V2I based environment perception for autonomous vehicles at intersections,, X. Duan et al., vol. 18, no. 7, pp. 1– 12, , 2021

78. A Kriging based spatiotemporal approach for traffic volume data imputation,, H. Yang et al., vol. 13, no. 4, e0195957, , 2018

79. Auto-scaling VNFs using machine learning to improve QoS and reduce cost, in, M. Tornatore and, T. Ahmed, M. Huynh, B. Mukherjee, S. Rahman, IEEE International Conference on Communications (ICC), pp. 1–6, , 2018

80. Data dissemination framework using low-rank approximation in edge networks,, J. Kwon and, H. Park, IEEE Access, vol. 12, pp. 1266–1279, 2023, , 2023

81. Fast composite optimization and statistical recovery in federated learning,, S. Luo and, Y. Bao, M. Liu, M. Crawshaw, in International Conference on Machine Learning (ICML), PMLR, 2022, pp. 1508–1536, , 2022

82. Bayesian temporal factorization for multidimensional time series prediction,, L. Sun, X. Chen and, IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 44, no. 9, pp. 4659–4673, , 2021

83. Bounds on the interchannel interference of OFDM in time-varying impairments,, Y. Li and, L. J. Cimini, IEEE Transactions on Communications, vol. 49, no. 3, pp. 401–404, , 2001

84. Gradient statistics aware power control for over-the-air federated learning,, M. Tao, N. Zhang and, IEEE Transactions on Wireless Communications, vol. 20, no. 8, pp. 5115–5128, , 2021

85. Study on network data feature importance for network node classification, in, J. Kwon, H. Park, D. Jung and, The Korean Institute of Communications and Information Sciences (KICS) Winter Conference, pp. 1245–1246, , 2020

86. Hourly day-ahead solar irradiance prediction using weather forecasts by LSTM,, Y. Niu, X. Qing and, vol. 148, pp. 461–468, , 2018

87. Reliability-aware network service provisioning in mobile edge-cloud networks,, X. Jia, M. Huang and, W. Liang, J. Li, IEEE Transactions on Parallel and Distributed Systems, vol. 31, no. 7, pp. 1545–1558, , 2020

88. SAFAA semiasynchronous protocol for fast federated learning with low overhead, L. He, R. Mao, C. Maple and, W. Wu, S. Jarvis, W. Lin, IEEE Transactions on Computers, vol. 70, no. 5, pp. 655–668, , 2020

89. The mnist database of handwritten digit images for machine learning research,, L. Deng, IEEE Signal Processing Magazine, vol. 29, no. 6, pp. 141–142, , 2012

90. TutorialComplexity analysis of singular value decomposition and its variants,, X. Li, Y. Cai, S. Wang and, arXiv preprint arXiv:1906.12085, , 2019

91. A lightweight ensemble spatiotemporal interpolation model for geospatial data,, F. Lu, S. Cheng, P. Peng and, vol. 34, no. 9, pp. 1849–1872, , 2020

92. FedCommUnderstanding communication protocols for edge-based federated learning, G. Cleland, D. Wu, B. Varghese, R. Ullah and, in 2022 IEEE/ACM 15th International Conference on Utility and Cloud Computing (UCC), IEEE, pp. 71–81, , 2022

93. Missing value imputationA review and analysis of the literature (2006–2017),, C.-F. Tsai, W.-C. Lin and, Artificial Intelligence Review, vol. 53, pp. 1487–1509, , 2020

94. Traffic data classification using machine learning algorithms in SDN networks,, D. Jung and, H. Park, J. Kwon, in 2020 International Conference on Information and Communication Technology Convergence (ICTC), IEEE, pp. 1031–1033, , 2020

95. A self-adaptive deep learningbased system for anomaly detection in 5G networks,, ´A. L. Perales G´omez, M. Gil P´erez and, L. Fern´andez Maim´o, F. J. Garc´ia Clemente, G. Mart´inez P´erez, IEEE Access, vol. 6, pp. 7700–7712, , 2018

96. Big vehicular traffic data miningTowards accident and congestion prevention, in, H. Al Najada and, I. Mahgoub, 2016 International Wireless Communications and Mobile Computing Conference (IWCMC), IEEE, pp. 256–261, , 2016

97. Edge AIOn-demand accelerating deep neural network inference via edge computing,, L. Zeng, E. Li, Z. Zhou and, X. Chen, IEEE Transactions on Wireless Communications, vol. 19, no. 1, pp. 447–457, , 2019

98. FedParaLow-rank hadamard product for communication-efficient federated learning, N. Hyeon-Woo, M. Ye-Bin and, T.-H. Oh, in International Conference on Learning Representations, , 2021

99. Federated learning over noisy channelsConvergence analysis and design examples,, X. Wei and, C. Shen, IEEE Transactions on Cognitive Communications and Networking, vol. 8, no. 2, pp. 1253–1268, , 2022

100. Practical approach for travel time estimation from point traffic detector data,, M. Hadi, L. Shen and, vol. 47, no. 5, pp. 526–535, , 2013

101. SOHO-FLA fast reconvergent intra-domain routing scheme using federated learning, K. Xu, W. Wang and, P. Cong, Y. Zhang, IEEE Network 2023, , 2023

102. Spatio-temporal stacked LSTM for temperature prediction in weather forecasting,, J. A. Suykens, Z. Karevan and, arXiv preprint arXiv:1811.06341, , 2018

103. Systematic network coding based reliable real-time multimedia streaming system,, H. Kim, S. Bai and, M. Kwon, H. Park, I. Zhang, in 2018 IEEE International Conference on Consumer Electronics (ICCE), IEEE, pp. 1–2, , 2018

104. ClusterGradAdaptive gradient compression by clustering in federated learning, in, L. Zhang, L. Cui, Y. Zhou and, X. Su, GLOBECOM 2020- 2020 IEEE Global Communications Conference, IEEE pp. 1–7, , 2020

105. Efficient Auto-Scaling approach in the telco cloud using self-learning algorithm, W. Hu and, P. Tang, L. Yang, F. Li, W. Zhou, in IEEE Global Communications Conference (GLOBECOM), pp. 1–6, , 2015

106. FhdnnCommunication efficient and robust federated learning for AIoT networks, in, J. Kang and, T. Rosing, K. Ergun, D. Nanjunda, R. Chandrasekaran, J. Lee, Proceedings of the 59th ACM/IEEE Design Automation Conference, pp. 37–42, , 2022

107. Towards end-to-end speech recognition with deep convolutional neural networks in, A. Courville, P. Brakel, M. Pezeshki, Y. Zhang, C. L. Y. Bengio and, S. Zhang, 17th Annual Conference of the Inter- national Speech Communication Association (INTERSPEECH 2016), pp. 410–414, , 2016

108. Wireless data acquisition for edge learningData-importance aware retransmission,, J. Zhang and, G. Zhu, D. Liu, Q. Zeng, K. Huang, IEEE transactions on wireless communications, vol. 20, no. 1, pp. 406– 420, , 2020

109. A convolutional recurrent autoencoder for spatio-temporal missing data imputation, A. Reza and, R. Amelia, in International Conferences on Artificial Intelligence, , 2019

110. Delay-complexity trade-off of random linear network coding in wireless broadcast,, Q. T. Sun and, R. Su, Z. Zhang, vol. 68, no. 9, pp. 5606–5618, , 2020

111. Efficient and reliable data dissemination over handover dynamics in v2i networks,, H. Park, J. Kwon and, in 2020 IEEE International Conference on Consumer Electronics (ICCE), IEEE, pp. 1–2, , 2020

112. Machine type communications in 3GPP networksPotential, challenges, and solutions,, T. Taleb and, A. Kunz, vol. 50, no. 3, pp. 178–184, , 2012

113. On model transmission strategies in federated learning with lossy communications,, Y. Zhou, L. Cui and, J. Liu, X. Su, IEEE Transactions on Parallel and Distributed Systems, vol. 34, no. 4, pp. 1173–1185, 2023, , 2023

114. Study on machine learning based abnormal behavior network node classification, in, D. Jung and, J. Kwon, H. Park, The Korean Institute of Communications and Information Sciences (KICS) Winter Conference, pp. 1232–1233, , 2020

115. Wind speed prediction with spatio–temporal correlationA deep learning approach,, X. Duan and, Q. Zhu, L. Zhu, Y. Liu, J. Chen, vol. 11, no. 4, p. 705, , 2018

116. D-LSTMShort-term road traffic speed prediction model based on GPS positioning data, H. Fu, X. Meng, L. Peng et al., IEEE Transactions on Intelligent Transportation Systems, vol. 23, no. 3, pp. 2021–2030, , 2020

117. Experimental results for artificial intelligence-based self-organized 5G networks,, M. Strufe and, W. Jiang, H. D. Schotten, in IEEE Annual International Symposium on Personal, Indoor, and Mobile Radio Communications (PIMRC), pp. 1–6, , 2017

118. Network coding based evolutionary network formation for dynamic wireless networks,, M. Kwon and, H. Park, IEEE Transactions on Mobile Computing, vol. 18, no. 6, pp. 1316–1329, , 2018

119. Singular vector and singular subspace distribution for the matrix denoising model,, X. DING and, K. WANG, Z. BAO, vol. 49, no. 1, pp. 370–392, , 2021

120. Content popularity prediction and caching for ICNA deep learning approach with SDN,, J. Zhang, W. Liu, Z. Liang, L. Peng and, J. Cai, IEEE Access, vol. 6, pp. 5075–5089, , 2018

121. Federated learning for 6G communicationsChallenges, methods, and future directions,, Y. Liu, D. Niyato, J. Kang, X. Yuan, Z. Xiong, X. Wang and, vol. 17, no. 9, pp. 105–118, , 2020

122. PFFNPeriodic feature-folding deep neural network for traffic condition forecasting,, K.-L. Tsui, Z. Zhang and, T. Wang, vol. 11, no. 2, pp. 3108–3120, 2023, , 2023

123. 6G-enabled edge AI for metaverseChallenges, methods, and future research directions,, L. Chang et al., vol. 7, no. 2, pp. 107–121, , 2022

124. Estimating packet loss rate in the access through application-level measurements, in, M. Meo, A. Servetti and, S. Basso, J. C. De Martin, Proceedings of the 2012 ACM SIGCOMM workshop on Measurements up the stack, pp. 7–12, , 2012

125. Preservation of the global knowledge by not-true distillation in federated learning,, Y. Shin, S.-Y. Yun, G. Lee, S. Bae and, M. Jeong, Advances in Neural Information Processing Systems, vol. 35, pp. 38 461– 38 474, , 2022

126. Throughput-delay analysis of random linear network coding for wireless broadcasting,, N. B. Shroff, A. Eryilmaz and, B. Swapna, IEEE Transactions on Information Theory, vol. 59, no. 10, pp. 6328–6341, , 2013

127. An adaptive information quantitybased broadcast protocol for safety services in VANET, T. Luo, Y. Hu et al., W. Wang, Mobile Information Systems, vol. 2016, , 2016

128. Compound TCP performance for industry 4.0 WiFiA cognitive federated learning approach,, S. Singh, S. R. Pokhrel and, IEEE Transactions on Industrial Informatics, vol. 17, no. 3, pp. 2143–2151, , 2020

129. Convolutional LSTM networkA machine learning approach for precipitation nowcasting, in, X. Shi et al., Conference on Neural Information Processing Systems, vol. 28, pp. 802–810, , 2015

130. A multi-attention tensor completion network for spatiotemporal traffic data imputation,, X. Wu, M. Xu, X. Wu, J. Fang and, vol. 9, no. 20, pp. 20 203–20 213, , 2022

131. FedselFederated SGD under local differential privacy with top-k dimension selection, in, Y. Cao, M. Yoshikawa and, R. Liu, H. Chen, Database Systems for Advanced Applications25th International Conference, DASFAA 2020, Jeju, South Korea,, 2020, Proceedings, Part I 25, Springer, 2020, pp. 485–501, , 2020

132. Implementation of network coding algorithm for single queue with single memory devices,, Y. Park and, S. Kim, S. Cho, H. Park, J. Kwon, in 2017 Ninth International Conference on Ubiquitous and Future Networks (ICUFN), IEEE pp. 627–629, , 2017

133. Machine learning in software Defined NetworksData collection and traffic classification, J. Dinis, P. Pinto, J. Tavares and, P. Amaral, H. S. Mamede, L. Bernardo, in IEEE International Conference on Network Protocols (ICNP), pp. 1–5, , 2016

134. A nonconvex low-rank tensor completion model for spatiotemporal traffic data imputation,, J. Yang and, X. Chen, L. Sun, Transportation Research Part CEmerging Technologies, vol. 117, p. 102 673, , 2020

135. Data-importance aware user scheduling for communication-efficient edge machine learning,, G. Zhu, J. Zhang and, K. Huang, D. Liu, IEEE Transactions on Cognitive Communications and Networking, vol. 7, no. 1, pp. 265–278, , 2021

136. Federated learning for wireless communicationsMotivation, opportunities, and challenges,, S. Niknam, H. S. Dhillon and, J. H. Reed, vol. 58, no. 6, pp. 46–51, , 2020

137. Missing data imputation for traffic congestion data based on joint matrix factorization,, X. Jia et al., Knowledge-Based Systems, vol. 225, p. 107 114, , 2021

138. Federated learning for the internet of thingsApplications, challenges, and opportunities,, A. S. Avestimehr, M. Zhang, L. Gao, C. He, T. Zhang, B. Krishnamachari and, IEEE Internet of Things Magazine, vol. 5, no. 1, pp. 24–29, , 2022

139. Sensing data supported traffic flow prediction via denoising schemes and ANNA comparison,, C. Xinqiang et al., vol. 20, no. 23, pp. 14 317–14 328, , 2020

140. Improving TCP performance over WiFi for internet of vehiclesA federated learning approach,, J. Choi, S. R. Pokhrel and, IEEE transactions on vehicular technology, vol. 69, no. 6, pp. 6798–6802, , 2020

141. Principal-component-analysis eigenvalue spectra from data with symmetry-breaking structure,, M. Rattray, D. C. Hoyle and, Physical Review E, vol. 69, no. 2, p. 026 124, , 2004

142. Wind speed time series imputation with a bidirectional gated recurrent unit (GRU) model, in, H. Tito-Chura and, V. Yana-Mamani, A. Flores, Proceedings of the Future Technologies Conference, Springer, vol. 2, 2022, pp. 445–458, , 2022

143. A joint learning and communications framework for federated learning over wireless networks,, M. Chen, S. Cui, Z. Yang, H. V. Poor and, W. Saad, C. Yin, IEEE Transactions on Wireless Communications, vol. 20, no. 1, pp. 269–283, , 2020

144. Efficient probabilistic information broadcast algorithm over random geometric topologies, in, R. Hu, 2015 IEEE Global Communications Conference (GLOBECOM), IEEE, pp. 1–6, , 2015

145. Exponentially small bounds on the expected optimum of the partition and subset sum problems,, G. S. Lueker, Random Structures & Algorithms, vol. 12, no. 1, pp. 51–62, , 1998

146. Adaptive gradient sparsification for efficient federated learningAn online learning approach,, S. Wang and, P. Han, K. K. Leung, in 2020 IEEE 40th international conference on distributed computing systems (ICDCS), IEEE, pp. 300–310, , 2020

147. EdenCommunication-efficient and robust distributed mean estimation for federated learning, in, R. B. Basat, G. Mendelson, Y. B. Itzhak and, M. Mitzenmacher, S. Vargaftik, A. Portnoy, International Conference on Machine Learning (ICML), PMLR, pp. 21 984–22 014, , 2022

148. Reliability-aware virtualized network function services provisioning in mobile edge computing,, Y. Ma and, M. Huang, X. Shen, H. Kan, W. Liang, IEEE Transactions on Mobile Computing, vol. 19, no. 11, pp. 2699–2713, , 2019

149. Dres-flDropout-resilient secure federated learning for non-iid clients via secret data sharing,, J. Shao, J. Zhang, Y. Sun, S. Li and, Advances in Neural Information Processing Systems, vol. 35, pp. 10 533–10 545, , 2022

150. Evaluation of ATSC 3.0 and 3GPP Rel-17 5G broadcasting systems for mobile handheld applications,, S.-K. Ahn et al., IEEE Transactions on Broadcasting, vol. 69, no. 2, pp. 338–356, , 2022

151. Deep learning-driven wireless communication for edge-cloud computingOpportunities and challenges,, X. Li and, H. Wu, Y. Deng, vol. 9, pp. 1–14, , 2020

152. Optimized network coding efficiency under QoS constraints in two-way relay networks with timeouts,, C. Skianis, E.-C. Davri, K. Kontovasilis and, Communications (ICC), 2015 IEEE International Conference on, pp. 6157–6162, , 2015

153. Missing value imputation for traffic-related time series data based on a multi-view learning method,, Y. Wang and, J. Zhang, B. Ran, L. Li, IEEE Transactions on Intelligent Transportation Systems, vol. 20, no. 8, pp. 2933–2943, , 2018

154. Wireless federated learning with hybrid local and centralized trainingA latency minimization design,, M. Dai, N. Huang, X. Shen, Y. Wu, T. Q. Quek and, IEEE Journal of Selected Topics in Signal Processing, vol. 17, no. 1, pp. 248–263, , 2022

155. Adjacent channel interference aware joint scheduling and power control for V2V broadcast communication,, A. Hisham, D. Yuan, E. G. Str¨om and, F. Br¨annstr¨om, vol. 22, no. 1, pp. 443–456, , 2020

156. Big data analytics, machine learning, and artificial intelligence in next-generation wireless networks,, G. P. Villardi, M. G. Kibria, O. Zhao, K. Nguyen, K. Ishizu and, F. Kojima, IEEE access, vol. 6, pp. 32 328– 32 338, , 2018

157. K nearest neighbours with mutual information for simultaneous classification and missing data imputation, P. J. Garc´ia-Laencina, M. Verleysen, A. R. Figueiras-Vidal and, J.-L. Sancho-G´omez, Neurocomputing, vol. 72, no. 7-9, pp. 1483–1493, , 2009

158. Understanding operational 5GA first measurement study on its coverage, performance and energy consumption,, D. Xu, X. Zhang et al., A. Zhou, in Proceedings of the Annual conference of the ACM Special Interest Group on Data Communication on the applications, technologies, architectures, and protocols for computer communication, pp. 479–494, , 2020

159. Virtual Network Function selection and chaining based on deep learning in SDN and NFV-enabled networks, in, D. Li, J. Pei, P. Hong and, IEEE International Conference on Communications Workshops (ICC Workshops), pp. 1–6, , 2018

160. Federated fullparameter tuning of billion-sized language models with communication cost under 18 Kilobytes,, Y. Li and, D. Chen, S. Deng, B. Qian, B. Ding, Z. Qin, arXiv preprint arXiv:2312.06353, 2023, , 2023

161. Known interference in the cellular downlinkA performance limiting factor or a source of green signal power?, A. K. Shukla, C. Masouros, T. Ratnarajah, M. Sellathurai, C. B. Papadias and, vol. 51, no. 10, pp. 162–171, , 2013

162. Road accidents detection, data collection and data analysis using V2X communication and edge/cloud computing,, J. Pannek, O. Chughtai, A. Qayyum and, K. A. Khaliq, A. Shahwani, Electronics, vol. 8, no. 8, p. 896, , 2019

163. Memory-augmented dynamic graph convolution networks for traffic data imputation with diverse missing patterns,, Y. Liang, Z. Zhao and, L. Sun, Transportation Research Part CEmerging Technologies, vol. 143, p. 103 826, , 2022

164. Road-speed profile for enhanced perception of traffic conditions in a partially connected vehicle environment,, M. A. S. Kamal et al., IEEE Transactions on Vehicular Technology, vol. 67, no. 8, pp. 6824–6837, , 2018

165. A modified inverse distance weighting method for interpolation in open public places based on Wi-Fi probe data,, L.-n. L. Li et al., vol. 2019, , 2019

166. Implementation of network-coded cooperation for energy efficient content distribution in 5G mobile small cells,, F. H. Fitzek, R. Torre, G. T. Nguyen and, S. Pandi, H. Salah, I. Leyva-Mayorga, IEEE Access, vol. 8, pp. 185 964–185 980, , 2020

167. 6g wireless communication systems: Applications, requirements, technologies, challenges, and research directions,, Y. M. Jang, M. Z. Chowdhury, M. Shahjalal, S. Ahmed and, vol. 1, pp. 957–975, , 2020

168. Proactive VNF provisioning with multi-timescale cloud resourcesFusing online learning and online optimization, in, C. Wu, X. Zhang, Z. Li and, F. C. M. Lau, IEEE Conference on Computer Communications (INFOCOM), pp. 1–9, , 2017

169. Using noise pollution data for traffic prediction in smart citiesExperiments based on LSTM recurrent neural networks,, F. M. Awan, R. Minerva and, N. Crespi, vol. 21, no. 18, pp. 20 722– 20 729, , 2021

170. Real-time data processing scheme using big data analytics in internet of things based smart transportation environment, F. Arif, M. Babar and, vol. 10, pp. 4167–4177, , 2019

171. Deep reinforcement learning-based dynamic resource management for mobile edge computing in industrial internet of things,, Y. Zhang, L. Zhao, Z. Liu, Y. Wu, Y. Chen, X. Chen and, IEEE Transactions on Industrial Informatics, vol. 17, no. 7, pp. 4925–4934, , 2020

172. Statistical learning-based dynamic retransmission mechanism for mission critical communicationAn edge-computing approach,, N. Shariati and, M. Abolhasan, J. Lipman, W. Ni, M. A. Raza, in 2020 IEEE 45th Conference on Local Computer Networks (LCN), IEEE, pp. 393– 396, , 2020

173. Truncated tensor schatten p-norm based approach for spatiotemporal traffic data imputation with complicated missing patterns,, T. Nie, J. Sun, G. Qin and, Transportation Research Part CEmerging Technologies, vol. 141, p. 103 737, , 2022

174. Performance analysis and threshold quantization of transformer differential protection under sampled value packets loss/delay,, Q. Jiang, L. Zhou and, J. Zhu, H. Peng, R. He, IEEE Access, vol. 7, pp. 55 698– 55 706, , 2019

175. ST-LBAGANSpatio-temporal learnable bidirectional attention generative adversarial networks for missing traffic data imputation, B. Yang et al., Knowledge-Based Systems, vol. 215, p. 106 705, , 2021

176. Coexistence analysis of H2H and M2M traffic in FiWi smart grid communications infrastructures based on multi-tier business models,, G. Joos, F. Aurzada, M. Levesque, M. Maier and, vol. 62, no. 11, pp. 3931–3942, , 2014

177. Software- Defined networks with mobile edge computing and caching for Smart citiesA big data deep reinforcement learning approach,, V. C. M. Leung and, Y. He, N. Zhao, F. R. Yu, H. Yin, IEEE Communications Magazine, vol. 55, no. 12, pp. 31–37, , 2017

178. Real-time spatiotemporal prediction and imputation of traffic status based on LSTM and Graph Laplacian regularized matrix factorization, J.-M. Yang, Z.-R. Peng and, L. Lin, Transportation Research Part CEmerging Technologies, vol. 129, p. 103 228, , 2021

179. Representing twentieth-century space–time climate variability. part IIDevelopment of 1901–96 monthly grids of terrestrial surface climate,, M. Hulme and, M. New, P. Jones, vol. 13, no. 13, pp. 2217–2238, , 2000

180. Missing traffic data imputation for artificial intelligence in intelligent transportation systemsReview of methods, limitations, and challenges, J. M.-Y. Lim and, R. Parthiban, R. C. K. Cheong, IEEE Access, vol. 11, pp. 34 080–34 093, 2023, , 2023

181. Fast inverse distance weightingbased spatiotemporal interpolationA web-based application of interpolating daily fine particulate matter PM2.5 in the contiguous US using parallel programming and k-d tree,, C. Yorke and, R. Piltner, L. Li, T. Losser, International Journal of Environmental Research and Public Health, vol. 11, no. 9, pp. 9101–9141, 2014., , 2014