🚀 Basics of Deep Learning

Welcome to the "General Deep Learning Learning" course! 🎉

📅 Course Details

• Dates: October. To be confirmed.
• Time: Wednesdays, 14:00 – 16:00
• Location: Virtual (Zoom link provided upon registration)
• Prerequisites: Basic Python & statistics and linear algebra knowledge recommended

---

🧠 What You'll Learn

This course covers core Deep Learning concepts and practical techniques to build robust models.

---

📚 Detailed Content

Basics of Neural Networks (NNs)

• Learning Objective and Loss Functions
• Optimization Problem Setup
• Gradients as Learning Signals
• Computational Graphs
• Automatic Differentiation (Backprop intuition)
• Basic Gradient Descent
• Integration: First Neural Network

Training Deep NNs

• Scaling the Network
• Optimization in Practice
• Failure Modes in Deep Networks (Vanishing/ exploding gradients, Saturating activations)
• Activation Functions
• Initialization
• Normalization
• Regularization and Generalization
• Diagnosing
• Integration: A Working Deep MLP

Convolutional NN and Computer Vision

• How to make a network see?
• Inductive biases, Convolution, and pooling
• Space invariance
• Padding, strides, and dilation

Representation Learning

• Analyzing SVM weights, detecting biases, and SHAP/LIME explanations

Advance Deep Learning topics

• Attention
• Transformers
• Foundation models
• AGI
• Fairness

---

🔗 Resources

🔗 Full Program: [Google Doc] 📂 Slides & Materials: [Sciebo]

👨‍💻👩‍💻 See you in class! Happy learning! 👨‍💻👩‍💻