Neural Networks and Deep Learning

Registration

Certification of attendance

Course Program and Lectures Schedule

1. The first module (20 hours) focuses on the theoretical foundations of neural networks.
2. The second module (20 hours) focuses on deep neural networks.
3. The third module (20 hours) covers more advanced topics and recent research trends.
4. The fourth module (30 hours) covers practical and implementation issues.

• Course program (pdf)
• Lecture schedule (pdf)

Part I: Theoretical Foundations

1. Introduction to neural computing
2. Hopfield networks
3. Unsupervised learning: PCA
4. Unsupervised learning: Self-Organizing Maps
5. Clustering algorithms - part 1
6. Clustering algorithms - part 2
7. Reinforcement learning - part 1
8. Reinforcement learning - part 2
9. Reinforcement learning - part 3
10. Supervised learning: The Perceptron
11. Supervised learning: Backpropagation
12. Supervised learning: Performance metrics
13. Radial Basis Function Networks

Part II: Deep Learning

1. Towards Deep Neural Networks (DNNs)
2. Autoencoders
3. Convolutional Neural Networks
4. CNNs for object classification
5. CNNs for object detection
6. CNNs for image segmentation
7. Deformable Convolutions
8. Recurrent Neural Networks - part 1
9. Recurrent Neural Networks - part 2
10. Transformers

Part III: Advanced Topics

1. Deep reinforcement learning - part 1
2. Deep reinforcement learning - part 2
3. Generative networks
4. Semi-supervised learning
5. Adversarial attacks and defenses
6. Real-world adversarial attacks
7. Explainable and interpretable AI
8. Anomaly detection and out-of-distribution generalization
9. Domain generalization and domain adaptation
10. Attention mechanisms in computer vision

Part IV: Implementation Issues

1. Implementing neural networks in C: Self Organizing Maps
2. Implementing neural networks in C: Q-learning
3. Implementing neural networks in C: Backpropagation
4. Programming frameworks for deep learning Examples
5. Modeling DNNs in Tensorflow and PyTorch Examples
6. Functional components in autonomous driving
7. The Apollo framework for autonomous driving
8. Simulators for autonomous driving
9. GPU programming in CUDA
10. Accelerating deep networks on GPGPUs
11. DNN optimization for embedded platforms
12. The NVIDIA TensorRT framework
13. Accelerating deep networks on FPGA
14. The Xilinx Deep Processing Unit

Exam

• Project rules
• Projects list
• Projects availability

Videos

1. Hopfield network demo
2. Kohonen network demo 1: 2D input space
3. Kohonen network demo 2: image mapping
4. Kohonen network demo 3: optimization
5. K-means online demo
6. Gaussian mix demo
7. Q-learning: The amoeba
8. Q-learning: cart-pole demo
9. Q-learning: grid world demo
10. Q-learning: crossroads demo
11. Q-learning: shooting moving targets
12. Q-learning: target camera tracking
13. Q-learning: car driving demo
14. Supervised learning: function approximator
15. Supervised learning: overfitting example
16. Supervised learning: simple classifier
17. Supervised learning: character classifier
18. Supervised learning: face classifier

Suggested readings

• Ian Goodfellow, Yoshua Bengio, Aaron Courville. Deep Learning, MIT Press, 2017.
• François Chollet. Deep Learning with Python, Manning, 2017.
• Richard S. Sutton and Andrew G. Barto. Reinforcement Learning: An Introduction, Second edition, The MIT Press, 2018.

• Giorgio Buttazzo, "Reti Neurali in grado di apprendere", Ithaca: Viaggio nella Scienza, XVI, pp. 195-208, 2020.

• Artificial Consciousness: Utopia or Real-Possibility? (2001)
  
• Coscienza Artificiale: Missione Impossibile? (2002)
  
• Artificial Consciousness: Hazardous Questions (and Answers) (2008)
  
• Rise of Artificial General Intelligence: Risks and Opportunities (2023)