Academic year: 2018-2019
Teaching coordinators: Jesse Read, Olivier Pietquin, Nikolaos Tziortziotis, Michalis Vazirgiannis
Due to recent breakthroughs, for example in the area of robotics (such as self-driving cars), and strategic game playing (such as Go), artificial intelligence is experiencing a revival, and is attracting increasing interest from major players in the academic and industrial scene.
An advanced artificial intelligence builds on the material presented from Machine Learning I, and thus allows greater freedom to explore advanced topics, since students will have a solid base in machine learning, which is a fundamental aspect to much of modern AI.
The following gives an approximate outline of the course. Lectures will cover the relevant theory, and labs will familiarize the students with these topics from a practical point of view. Two of the lab assignments will be graded, and a team project on reinforcement learning will form a major component of the grade - where the goal is to developing and deploy an agent in an environment and write a report analyzing the results.
Syllabus
- Introduction & Foundations: Probabilistic Reasoning and Decision Making
- Multi-output and Structured-Output Prediction
- Search & Optimization
- Reinforcement Learning
- Bayesian Filtering; Recurrent Neural Networks
- Deep Neural Networks
- Deep Reinforcement Learning
- Self Play; Generative Adversarial Networks
- Learning from Concept-Drifting Data Streams
Source : https://moodle.polytechnique.fr/course/info.php?name=INF581-2018