This graduate-level course explores the rapidly evolving field of deep learning through a statistical lens. While covering the foundational architectures and techniques of modern neural networks, the emphasis will be on understanding these methods as powerful tools for statistical modeling, inference, and prediction. We will delve into topics such as model uncertainty, generalization, regularization, probabilistic deep learning, and the connections between deep learning and established statistical concepts. Students will gain both a theoretical understanding and practical experience in implementing and evaluating deep learning models for complex data.

The goal of this course is to provide students with a comprehensive understanding of deep learning from a statistical perspective. By the end of the course, students should be able to: - Understand the foundations of deep learning and its statistical implications. - Implement and evaluate various deep learning architectures. - Apply probabilistic models and inference techniques in deep learning. - Analyze model uncertainty and generalization in deep learning. - Connect deep learning methods with traditional statistical approaches.

- Linear algebra, calculus, probability, and statistics. - Programming experience, preferably in Python, e.g., NumPy and PyTorch.

The instructor will provide slides and the lectures are mainly derived from the following books: 1. Goodfellow, I., Bengio, Y., & Courville, A. (2016). Deep learning. (DL) 2. Zhang, A., Lipton, Z. C., Li, M., & Smola, A. J. (2023). Dive into deep learning (http://d2l.ai/index.html). (D2L)