References

References

1. T. J. O'Shea and J. Hoydis, An Introduction to Deep Learning for the Physical Layer, IEEE Transactions on Cognitive Communications and Networking, vol. 3, no. 4, pp. 563--575, 2017

   The seminal paper introducing the end-to-end autoencoder framework for communication systems. Demonstrates that a neural network can jointly learn transmitter and receiver functions, discovering known optimal constellations and novel arrangements for non-standard channels.

2. K. Gregor and Y. LeCun, Learning Fast Approximations of Sparse Coding, Proc. 27th International Conference on Machine Learning (ICML), pp. 399--406, 2010

   Introduces LISTA (Learned ISTA), the foundational deep unfolding architecture that converts ISTA iterations into trainable neural network layers. Shows that 10 LISTA layers can match hundreds of ISTA iterations.

3. X. Chen, J. Liu, Z. Wang, and W. Yin, Theoretical Linear Convergence of Unfolded ISTA and Its Practical Weights and Thresholds, Advances in Neural Information Processing Systems (NeurIPS), vol. 31, pp. 9061--9071, 2018

   Provides the first theoretical convergence guarantee for LISTA, proving linear (geometric) convergence in the number of layers under mild conditions. Establishes that the learned weights and thresholds have a principled interpretation.

4. B. McMahan, E. Moore, D. Ramage, S. Hampson, and B. A. y Arcas, Communication-Efficient Learning of Deep Networks from Decentralized Data, Proc. 20th International Conference on Artificial Intelligence and Statistics (AISTATS), pp. 1273--1282, 2017

   Introduces the Federated Averaging (FedAvg) algorithm for training neural networks across multiple clients without sharing raw data. Demonstrates convergence on non-IID data and establishes the foundations of federated learning.

5. T. Li, A. K. Sahu, M. Zaheer, M. Sanjabi, A. Talwalkar, and V. Smith, Federated Optimization in Heterogeneous Networks, Proc. Conference on Machine Learning and Systems (MLSys), pp. 429--450, 2020

   Provides a convergence analysis of FedAvg under heterogeneous (non-IID) data distributions and proposes FedProx, which adds a proximal term to the local objective to reduce client drift. Quantifies the heterogeneity bias.

6. N. Shlezinger, J. Whang, Y. C. Eldar, and A. G. Dimakis, Model-Based Deep Learning, Proceedings of the IEEE, vol. 111, no. 5, pp. 465--499, 2023

   Comprehensive tutorial on model-based deep learning, covering deep unfolding, DNN-aided algorithms, and learned optimisation. Provides a unified framework and design guidelines for integrating domain knowledge into neural network architectures.

7. V. Mnih, K. Kavukcuoglu, D. Silver, A. A. Rusu, J. Veness, M. G. Bellemare, A. Graves, M. Riedmiller, A. K. Fidjeland, G. Ostrovski, S. Petersen, C. Beattie, A. Sadik, I. Antonoglou, H. King, D. Kumaran, D. Wierstra, S. Legg, and D. Hassabis, Human-Level Control Through Deep Reinforcement Learning, Nature, vol. 518, pp. 529--533, 2015

   Introduces Deep Q-Networks (DQN) with experience replay and target networks. While not specific to wireless, this paper established the deep RL framework that underpins most RL-based wireless resource allocation research.

8. H. He, C.-K. Wen, S. Jin, and G. Y. Li, Model-Driven Deep Learning for MIMO Detection, IEEE Transactions on Signal Processing, vol. 68, pp. 1702--1715, 2019

   Applies deep unfolding to MIMO detection by unrolling projected gradient descent (DetNet) and ADMM. Demonstrates that 10 unfolded layers with learned step sizes achieve near-ML detection performance at a fraction of the computational cost.

9. Y. S. Nasir and D. Guo, Multi-Agent Deep Reinforcement Learning for Dynamic Power Allocation in Wireless Networks, IEEE Journal on Selected Areas in Communications, vol. 37, no. 10, pp. 2239--2250, 2019

   Applies multi-agent deep RL to distributed power control in interference channels. Each transmitter is an independent agent that learns a power control policy from local observations, demonstrating that decentralised learning can approach centralised performance.

10. M. Chen, Z. Yang, W. Saad, C. Yin, H. V. Poor, and S. Cui, A Joint Learning and Communications Framework for Federated Learning Over Wireless Networks, IEEE Transactions on Wireless Communications, vol. 20, no. 1, pp. 269--283, 2022

    Develops a joint optimisation framework for federated learning over wireless, including resource allocation for model upload, client scheduling, and over-the-air aggregation. Bridges the ML and wireless communications perspectives.

Further Reading

• Deep learning for the physical layer (comprehensive survey)

  T. Wang, C.-K. Wen, H. Wang, F. Gao, T. Jiang, and S. Jin, 'Deep Learning for Wireless Physical Layer: Opportunities and Challenges,' China Communications, vol. 14, no. 11, pp. 92--111, Nov. 2017

  One of the earliest comprehensive surveys of deep learning applications at the physical layer, covering channel estimation, signal detection, coding, and end-to-end learning. Provides useful context for the rapid development of the field.

• Reinforcement learning for wireless networks (survey)

  F. B. Mismar, B. L. Evans, and A. Alkhateeb, 'Deep Reinforcement Learning for 5G Networks: Joint Beamforming, Power Control, and Interference Coordination,' IEEE Trans. Commun., vol. 68, no. 3, pp. 1581--1592, Mar. 2020

  Demonstrates the application of deep RL to joint beamforming and power control in mmWave 5G networks, showing how RL agents can handle the complex optimisation landscape of modern cellular systems.

• Federated learning over wireless networks

  K. Yang, T. Jiang, Y. Shi, and Z. Ding, 'Federated Learning via Over-the-Air Computation,' IEEE Trans. Wireless Commun., vol. 19, no. 3, pp. 2022--2035, Mar. 2020

  Develops the over-the-air aggregation concept for federated learning, where the wireless MAC channel is exploited to compute the model average for free. A key paper connecting FL theory with wireless physical-layer design.

• Algorithm unrolling for signal processing (tutorial)

  V. Monga, Y. Li, and Y. C. Eldar, 'Algorithm Unrolling: Interpretable, Efficient Deep Learning for Signal Processing,' IEEE Signal Processing Magazine, vol. 38, no. 2, pp. 18--44, Mar. 2021

  An accessible tutorial on the deep unfolding methodology, covering the theoretical foundations, design principles, and applications across signal processing. Excellent companion reading for the LISTA material in Section 31.2.

• Semantic communications and joint source-channel coding

  H. Xie, Z. Qin, G. Y. Li, and B.-H. Juang, 'Deep Learning Enabled Semantic Communication Systems,' IEEE Trans. Signal Processing, vol. 69, pp. 2663--2675, 2021

  Extends the autoencoder paradigm to semantic communications, where the goal is to transmit the meaning of data (text, images) rather than exact bit sequences. Represents the frontier of ML-driven communication system design.