Dipartimento di Ingegneria dell'Informazione, delle Infrastrutture e dell'Energia Sostenibile - Università Mediterranea di Reggio Calabria - Didattica

Questo sito utilizza cookie tecnici e di terze parti. Se vuoi saperne di più o negare il consenso consulta l'informativa sulla privacy. Proseguendo la navigazione o cliccando su "Chiudi" acconsenti all'uso dei cookie. Chiudi

Neural Engineering

Corso

Ingegneria Elettrica ed Elettronica LM-29

Curriculum

Electrical and Electronic Engineering

Orientamento

Orientamento unico

Anno Accademico

2022/2023

Crediti

Settore Scientifico Disciplinare

ING-IND/31

Anno

Secondo anno

Unità temporale

Secondo semestre

Ore aula

Attività formativa

Attività formative a scelta dello studente (art.10, comma 5, lettera a)

Canale unico

Docente

Nadia Mammone

Obiettivi

KNOWLEDGE AND UNDERSTANDING
The course aims at completing the master's student training by extending it with knowledge and skills in the field of artificial intelligence (with specific attention to deep learning methods), focusing the attention on the topics more related to neuroscience. Applications to electroencephalographic (EEG) signal processing are proposed both for the discrimination of brain states and for Brain Computer Interface (BCI). The student, independently, will develop a project according to the suggestions provided by the teacher.
ABILITY TO APPLY KNOWLEDGE AND UNDERSTANDING
The course aims at providing the student with skills on the development of algorithms based on Deep Learning methods, especially for applications in the field of neuroscience. This objective is also pursued through laboratory experiments and training sessions during which the student will have the opportunity to design his/her own experiments, acquire the electroencephalographic signals and process them.
INDEPENDENT JUDGMENT
At the end of the course, and in particular upon exam completion, the student will be able to: define paradigms and experiments for acquiring bio-signals for the achievement of the objectives of the study; design algorithms for processing the acquired signals in order to achieve the goal of the study. The student will be able to rely on the methods covered during the course as well as on his/her own ability to independently explore additional methods found in the literature, thanks to an acquired ability to handle the fundamental concepts of deep learning.
COMMUNICATION SKILLS
Thanks to a continuous student-teacher interaction, the course will focus on the acquisition of the language peculiar of artificial intelligence and computational neuroscience, also allowing the student to refine his/her own ability to express themselves through an appropriate technical language.
LEARNING SKILLS
Upon exam completion, the student will have acquired the ability to design algorithms based on artificial neural networks, both shallow and deep. He/she will be able to: define paradigms and experiments for the acquisition of electroencephalographic signals (EEG), according to the objectives of the study; acquire such signals; process them using Deep Learning algorithms developed ad-hoc; integrate such algorithms with platforms for interfacing them in real time with the EEG acquisition system. Conversely, the practical application of theoretical concepts will reinforce their understanding; the student will thus develop a theoretical-practical approach to engineering topics.

Programma

FUNDAMENTALS OF NEURAL ENGINEERING (0.5 CFU)
Fundamentals about neural engineering. Overview about Neural Engineering Applications. Need for a novel perspective in model-based approaches.
Introduction to the electric fields of the brain and to EEG. Fundamentals of EEG rhythms generation. EEG recording and acquisition. Normal vs. abnormal EEG patterns. Neurological disorders and their effects on brain waves (Alzheimer’s Disease, epilepsy, stroke, Parkinson’s Disease, etc). Brain Computer Interfaces (BCI) and EEG-based BCI. Application of Artificial Intelligence to EEG signals.

ARTIFICIAL NEURAL NETWORKS (2 CFU)
General properties of neural processing systems. Biological neuron model. McCulloch-Pitts artificial neuron. Gradient Descent. Nonlinearities: sigmoidal, hyperbolic tangent, ReLu activation functions.
Learning process. Error-Correction. Widrow-Hopf Rule. Hebbian Learning. Competitive Learning. Supervised and Unsupervised learning. Reinforcement Learning. Statistical Nature of the Learning Process.
Network architectures: feedforward and feedback models. Competitive and Self-Organizing models. Knowledge representation. Visualization of processes in Neural Networks.

Perceptrons. Multilayer Perceptrons. Radial-Basis Function Networks. Recurrent Networks. Self-Organizing Maps. Information-Theoretic Models. Temporal processing with neural networks.

DEEP LEARNING (1.5 CFU)
Deep vs shallow models. Convolutional Neural Networks. Visualizing and Understanding Convolutional Networks. Basic concepts of Explainable Machine Learning methods, Stacked AutoEncoders and Generative Adversarial Networks (GAN).

EEG SIGNAL PROCESSING (1 CFU)
Fundamentals of EEG signal processing. Analysis of EEG signals in the time domain. Spectral and time-frequency analysis of EEGs. Dynamical analysis and chaos. Entropic analysis. Different types of complexity. Graph theory applied to EEG. Principal Component Analysis and Independent Component Analysis. Common Spatial Patterns.
Neural Networks-based EEG signal processing. Classification of brain states through Neural Networks.

LABORATORY EXPERIMENTS (1 CFU)
Use of Matlab Deep Learning toolbox and/or Python.
Designing paradigms for offline and online EEG processing applications through EEGlab. Set up of an EEG recording system.
EEG recording and processing. Acquisition of bio-signals meant for the final project development.

Testi docente

José C. Principe, Neural and Adaptive Systems: Fundamentals Through Simulations, Wiley
Ian Goodfellow, Yoshua Bengio and Aaron Courville, “Deep Learning” (www.deeplearningbook.org), An MIT Press book
Zhang, A., Lipton, Z. C., Li, M., & Smola, A. J. (2021). Dive into deep learning. arXiv preprint arXiv:2106.11342 (https://d2l.ai/index.html)
Paul L. Nunez and Ramesh Srinivasan, Electric fields of the brain - the Neurophysics of EEG (second edition), Oxford University Press
Sani-Chambers, EEG Signal Processing, IEEE- Wiley
Simon Haykin, Neural Networks, IEEE Press

Erogazione tradizionale

Sì

Erogazione a distanza

Frequenza obbligatoria

Valutazione prova scritta

Valutazione prova orale

Sì

Valutazione test attitudinale

Valutazione progetto

Sì

Valutazione tirocinio

Valutazione in itinere