Avviso Seminario: "From Chaos to Artificial Intelligence in Wireless Communications" (27-30 Ottobre)

Si avvisano gli studenti e le studentesse interessati/e che nei giorni 27, 28, 29 e 30 Ottobre si terrà il seminario dal titolo "From Chaos to Artificial Intelligence in Wireless Communications".

Il seminario avrà una durata complessiva di 10 ore ed è strutturato secondo il seguente programma:

27 Ottobre: Module 1 (3 hours) - Fundamentals of Chaos Theory and Nonlinear Dynamical Systems

• Introduction to dynamical systems
• Linearity and nonlinearity
• Sensitivity to initial conditions
• Attractors, bifurcations, and complexity
• Main chaotic maps: Logistic Map, Bernoulli Map, Tent Map
• Statistical properties of chaotic signals
• Numerical examples and simulations
• Objective: Provide students with the fundamental knowledge needed to understand how fully deterministic systems can generate sequences that appear random, and how these properties can be exploited in telecommunication systems.

28 Ottobre: Module 2 (3 hours) - Chaos-Based Communications: Principles, Techniques, and Applications

• Introduction to Chaos-Based Communications
• Motivation and application scenarios
• Advantages of chaotic signals: robustness to interference, noise-like properties, physical-layer signal security
• Chaos Shift Keying (CSK) e Differential Chaos Shift Keying (DCSK)
• Multi-carrier techniques
• Transmission over AWGN and Rayleigh channels
• Synchronization and decoding issues
• Objective: This module will cover both the theoretical aspects and simple implementation and performance evaluation examples.

29 Ottobre: Module 3 (2 hours) - Deep Learning for Wireless Communications and Intelligent Receivers

• Limitations of traditional receivers
• Data-driven receivers
• Introduction to neural networks for the physical layer
• CNNs for sequence processing
• LSTMs and recurrent neural networks
• Attention mechanisms
• Compact models for embedded devices
• Application of AI to chaotic signal decoding
• Case Study: A case study from recent research activities will be presented, focusing on the development of lightweight neural receivers for chaotic communications. It will illustrate how Deep Learning and Attention techniques can significantly improve decoding performance in scenarios characterized by low signal-to-noise ratios and fading channels.

30 Ottobre: Module 4 (2 hours) - Research Frontiers: Towards Intelligent Communications of the Future

• Edge AI and wireless communications
• Neural network model compression and pruning
• Jamming-resilient communications
• AI-native communications
• Semantic communications
• Evolution towards 6G networks
• Open challenges and research opportunities