Igel, C., Riedmiller, M., Reinforcement learning in a Nutshell. ESANN'2007 Proceedings - European Symposium on Artificial Neural Networks, Bruges (Belgium), 277–288, 2007.
8. Ivanov, S. and D’yakonov, A., Modern Deep Reinforcement Learning Algorithms, arXiv preprint arXiv:1906.10025, 1–56, 2019.
9. Le Pham, T., Layek, A., Vien, N., Chung, T.C., Deep reinforcement learning algorithms for steering an underactuated ship in: 2017 IEEE International Conference on Multisensor Fusion and Integration for Intelligent Systems (MFI 2017), 602–607, 2017, 10.1109/MFI.2017.8170388.
10. Li, M.-J., Li, A.-H., Huang, Y.-J., Chu, S.-I., Implementation of Deep Reinforcement Learning. ICISS 2019: Proceedings of the 2019 2nd International Conference on Information Science and Systems, pp. 232–236, 2019, 10.1145/3322645.3322693.
11. Liu, Q., Zhai, J.-W., Zhang, Z.-Z., Zhong, S., Zhou, Q., Zhang, P., Xu, J., A Survey on Deep Reinforcement Learning. Jisuanji Xuebao/Chin. J. Comput., 41, 1–27, 2018, 10.11897/SP.J.1016 .2018. 00001.
12. Mnih, V., Kavukcuoglu, K., Silver, D., Rusu, A., Veness, J., Bellemare, M., Graves, A., Riedmiller, M., Fidjeland, A., Ostrovski, G., Petersen, S., Beattie, C., Sadik, A., Antonoglou, I., King, H., Kumaran, D., Wierstra, D., Legg, S., Hassabis, D., Human-level control through deep reinforcement learning. Nature, 518, 529–33, 2015, 10.1038/nature14236.
13. Mosavi, A., Ghamisi, P., Faghan, Y., Duan, P., Band, S., Comprehensive Review of Deep Reinforcement Learning Methods and Applications in Economics, Mathematics, 8, 1–42, 2020, 10.20944/preprints 202003.0309.v1.
14. Mousavi, S., Schukat, M., Howley, E., Deep Reinforcement Learning: An Overview, in: Lecture Notes in Networks and Systems, pp. 426–440, 2018, 10.1007/978-3-319-56991-8_32.
15. Nguyen, C., Dinh Thai, H., Gong, S., Niyato, D., Wang, P., Liang, Y.-C., Kim, D.I., Applications of Deep Reinforcement Learning in Communications and Networking: A Survey. IEEE Commun. Surv. Tutorials, 21, 4, 1–1, 2019, 10.1109/COMST.2019.2916583.
16. Sangeetha, S.K.B. and Ananthajothi, K., Machine Learning Tools for Digital Pathology—The Next Big Wave in Medical Science. Solid State Technol., 63, 3732–3749, 2020.
17. Santhya, R., Latha, S., Balamurugan, S., Charanyaa, S., Further investigations on strategies developed for efficient discovery of matching dependencies. Int. J. Innov. Res. Comput. Commun. Eng. (An ISO 3297:2007 Certified Organization), 3, 18998–19004, 2014.
18. Tan, F. and Yan, P., Deep Reinforcement Learning: From Q-Learning to Deep Q-Learning, Springer, Cham, Guangzhou, China, pp. 475–483, 2017, 10.1007/978-3-319-70093-9_50.
19. Xiliang, C., Cao, L., Li, C.-X., Xu, Z.-X., Lai, J., Ensemble Network Architecture for Deep Reinforcement Learning. Math. Probl. Eng., 2018, 1–6, 2018, 10.1155/2018/2129393.
20. Li, Y., Deep Reinforcement Learning: An Overview, arXiv:1701.07274v6, 2017.
1 * Corresponding author: [email protected]
2
Impact of AI in 5G Wireless Technologies and Communication Systems
A. Sivasundari* and K. Ananthajothi†
Department of Computer Science and Engineering, Misrimal Navajee Munoth Jain Engineering College, Chennai, India
Abstract
4G networks (with Internet Protocol or IP, telecommunications and reaction-based connectivity) have managed the network architecture. They have evolved and are now accessible in a multitude of ways, including advanced learning and deep learning. 5G is flexible and responsive and will establish the need for integrated real time decision-making. As the rollout has begun across the globe, recent technical and architectural developments in 5G networks have proved their value. In various fields of classification, recognition and automation, AI has already proved its efficacy with greater precision. The integration of artificial intelligence with internet-connected computers and superfast 5G wireless networks opens up possibilities around the globe and even in outer space. In this section, we offer an in-depth overview of the Artificial Intelligence implementation of 5G wireless communication systems. The focus of this research is in this context, to examine the application of AI and 5G in warehouse building and to discuss the role and difficulties faced, and to highlight suggestions for future studies on integrating Advanced AI in 5G wireless communications.
Keywords: Artificial intelligence, 5G, deep learning, machine learning, mobile networks, wireless communication
2.1 Introduction
Although 5G provides low latency and very high speed support capabilities (e.g., eMBB), a wide number of devices (e.g., mMTC), a heterogeneous mix of traffic types from a diverse and challenging suite of applications (e.g., URLLC), AI is complemented by observing from specific environments to provide independent reach of operation, turning 5G into a data-driven adaptive real-time network [13]. AI is used for 5G system modeling, automation of core network (e.g. provisioning, scheduling, prediction of faults, protection, fraud detection), distributed computing, reduction of operating costs, and improvement of both service quality and customer evolves on chatbots, recommendation systems, and strategies such as automated processes. In addition, AI is used across all layers, from the disaggregated radio access layer (5G RAN) to the distributed cloud layer (5G Edge/Core) to the integrated access backhaul to fine tune performance [5].
AI is used for the 5G distributed cloud layer to optimize device resource usage, autoscaling, identification of anomalies, predictive analytics, prescriptive policies, and so on. In addition, the 5G distributed cloud layer offers acceleration technologies to enable federated and distributed learning for AI workloads [19].
Figure 2.1 Growth of 5G Connections worldwide.
The growth of mobile 5G worldwide and its market overview are shown in Figure 2.1 and Figure 2.2. To support a multitude of emerging technologies (e.g., AR/VR, Industrial IoT, autonomous vehicles, drones, Industry 4.0 projects, Smart Cities, Smart Ports), AI is also used for customer service management and business support systems. Several learning methods are used to train data models based on target use cases, such as supervised learning, unsupervised learning, reinforcement learning, federated learning, distributed learning, transfer learning, and deep learning based on algorithms such as Convolutional Neural Networks (CNNs), Recurrent Neural Networks (RNNs) [9].
Figure 2.2 5G Market analysis.
Source: Press Release, Investor Relation Presentation, Annual Report, Expert Interview, and Markets and Markets Analysis
AI needs a mixture of AI, localized AI, and end-to-end AI at the system stage. In individual network modules, device-level AI is used to solve self-contained problems where no data needs to be transferred onto the network. Where AI is extended to one network domain or cross-network domains, localized AI allows data to be passed on to the network, but is limited to a local network domain, such as at the RAN or fronthaul. End-to-end AI is where the whole network needs to be accessible to the network, and where it needs to gather data and information from various network domains in order to implement AI properly. Slice management and network service assurance can provide examples of end-to-end AI [4, 8].
5G and AI are some of the world’s most groundbreaking innovations in hundreds