Deep Learning Spring 2019 Prof. Gilles Louppe g.louppe@uliege.be 1 / 12 Logistics This course is given by: Theory: Prof. Gilles Louppe (g.louppe@uliege.be) Projects and guidance: Joeri Hermans (joeri.hermans@doct.uliege.be) Matthia Sabatelli (m.sabatelli@uliege.be) Antoine Wehenkel (antoine.wehenkel@uliege.be) Feel free to contact any of us for help! 2 / 12 Lectures Theoretical lectures Tutorials Q&A sessions 3 / 12 Materials Slides are available at github.com/glouppe/info8010-deep-learning. In HTML and in PDFs. Posted online the day before the lesson (hopefully). Some lessons are partially adapted from "EE-559 Deep Learning" by Francois Fleuret at EPFL. 4 / 12 Textbook None! 5 / 12 Resources Awesome Deep Learning Awesome Deep Learning papers 6 / 12 AI at ULiège This course is part of the many other courses available at ULiège and related to AI, including: INFO8006: Introduction to Arti cial Intelligence ELEN0062: Introduction to Machine Learning INFO8010: Deep Learning ← you are there INFO8003: Optimal decision making for complex problems INFO8004: Advanced Machine Learning INFO0948: Introduction to Intelligent Robotics INFO0049: Knowledge representation ELEN0016: Computer vision DROI8031: Introduction to the law of robots 7 / 12 Outline (Tentative and subject to change!) Lecture 1: Fundamentals of machine learning Lecture 2: Neural networks Lecture 3: Convolutional neural networks Lecture 4: Training neural networks Lecture 5: Recurrent neural networks Lecture 6: Auto-encoders and generative models Lecture 7: Generative adversarial networks Lecture 8: Uncertainty Lecture 9: Adversarial attacks and defenses 8 / 12 Philosophy Thorough and detailed Understand the foundations and the landscape of deep learning. Be able to write from scratch, debug and run (some) deep learning algorithms. State-of-the-art Introduction to materials new from research (≤ 5 years old). Understand some of the open questions and challenges in the eld. Practical Fu