About

This course provides a comprehensive exploration of modern deep learning techniques, from foundational concepts to advanced topics.

Syllabus

1. MLP, Backpropagation
2. Optimization, Regularization
3. Initialization, Normalization, CNN
4. Introduction to Natural Language Processing, Word Embeddings
5. RNN, LSTM, Attention, Transformer
6. Classification, Object Detection
7. Segmentation
8. Multi-armed Bandits, Bellman Equations, Monte Carlo Methods, TD Learning, Q-Learning
9. Planning and Double Learning, Policy Improvement, Policy Gradient
10. Autoregression, VAE, GAN
11. Diffusion, Flow Matching
12. LLMs, Fine-tuning, LoRA, RAG, Agents
13. Multi-modal Models, CLIP, Qwen-VL
14. Quantization, Pruning, Distillation, KV-Cache, Flash Attention

Labworks

6 homeworks on practical implementation of Deep Learning models.

Grading

6 homeworks give 70 points in total + an exam for 30 points. Final score: min(round(#points/10), 10).

Prerequisites

• Probability Theory + Statistics
• Machine Learning
• Python