Description

Theory
2

Laboratory
2

Instructors

Luís Filipe Coelho

Contents

- Fundamentals of Signal Representation and Processing
- Artificial Intelligence applied to signals
- Advanced AI methods for signals
- Applications in Engineering
- AI in Embedded Systems and Edge Computing
- Design and Ethical Considerations

Learning Outcomes

The primary purpose of this unit is to empower students to design, implement, and evaluate AI-driven solutions for signal processing across diverse domains.
1. Signal Representation: Master time-frequency representations, including non-parametric (Mel, Cepstrum), parametric (Burg, Yule-Walker), and Wavelet models.
2. Data Management: Plan and curate signal datasets, ensuring quality for model training.
3. Model Development: Apply classification, clustering, and multimodal techniques (combining audio, video, and physiological signals).
4. Advanced Processing: Develop AI-based adaptive filters (RNN, LSTM), hybrid models (Kalman/DL), and generative models for signal synthesis.
5. Source Separation: Implement algorithms to isolate specific signal components.
6. Ethical Deployment: Understand Explainable AI (XAI) and address ethics, reliability, and safety in critical signal applications.