이 강의에서는 딥러닝의 이론적 정의와 실제 구현의 간극을 매우는 것을 목표로 합니다. 학생들은 DNN을 설계하는 것은 물론, 이들이 실제 하드웨어에서 어떻게 실현되는지와 최적화하는지에 대해 배우게 됩니다.
강의 내용을 깊이있게 이해할 수 있기 위하여, 인공지능 및 컴퓨터 아키텍처에 대한 기초 지식을 필요로 합니다. 본 강의의 최우선 목표는 아래와 같습니다.
1. Understand and explain key deep learning concepts and architectures.
2. Implement deep learning models
3. Optimize model performance for efficient training and inference.
4. Apply learned skills to real-world projects and research settings.

Artificial Intelligence / Computer Architecture

Deep Learning, Ian Goodfellow, Yoshua Bengio, Aaron Courville, MIT Press, 2016, ISBN: 978-0262035613
Dive into Deep Learning, Aston Zhang, Zachary C. Lipton, Mu Li, and Alexander J. Smola (online free textbook)

1. Introduction & Neural Network Basics / Matrix Multiplication
2. Convolution Neural Networks / High-rank Tensor Operations
3. Forward & Backward Propagation / Differentiable Operation & Operator Graph Building
4. Automatic Differentiation & Optimization / PyTorch Basics & Build Training Pipeline
5. ML Frameworks / Distributed Data Parallel with NVIDIA DALI
6. Computing System for DNNs / Profiling & Performance Analysis
7. Computing System for DNNs & Compilation / Design Custom Operator on PyTorch
8. Midterm exam
9. CUDA & Triton Language / Accelerating Custom Operators
10. Transformer and Large Language Models / LLM Frameworks -1
11. Computing System for LLMs / LLM Frameworks -2
12. Advanced Generative Models & Multimodal Models / Fine-tuning with LoRA
13. Distributed Training / Final Project -1
14. Model Optimization / Final Project -2
15. Reserved / Final Project -3
16. Final exam / Project Demonstration

Lecture, Hands-on Implementation, and Team Projects

Active participation in lectures and coding sessions is crucial.
Students are encouraged to explore advanced topics for their final projects.
S/U policy will not be allowed