• Empirical Risk Minimization
• Generative learning (MLE)
• Bayesian learning
• Generalization theory (PAC learning)
• Linear models (Perceptron algorithms, SVM,..)
• EM Algorithm
• Bias-Variance trade-off
• Model selection and validation methods
• Performance metrics
• VC dimension
• Deep learning and generalization

• Recap of Machine Learning
• Backpropagation (+implicit differentiation)
• Convolutional Neural Networks (+equivariances)
• Training Deep Models (initialization, normalization, residual)
• Regularization Methods for NNs
• Stochastic Gradient Descent (SGD)
• Adversarial patterns. Robust learning approaches
• Adaptive SGD methods
• Learning Representations I: Word Vectors, Metric Learning
• Learning Representations II: Unsupervised Representation Learning, VAE
• Graph Neural Networks
• Self-Attention, Transformers, ViT, LLMs
• TBA (Recurrent NNs + State Space Models / GAN + Diffusion / Fairness)

• Motivation (successes, AGI, human feedback, history)
• Multi-armed bandit problems (stochastic, contextual)
• Solving MDPs 1: (Bellman equations, Value iteration)
• Solving MDPs 2: (Contraction, Policy iteration)
• Temporal difference learning 1: (TD(0), Sarsa, Q-learning)
• Temporal difference learning 2: (n-step,Double-Q, DQN)
• Policy gradient methods 1: (Tabular)
• Policy gradient methods 2: (Variance reduction, Neural)
• Combining learning and planning (AlphaZero, muZero)
• Exploration in RL
• Multi-agent RL (cooperative vs. adversarial)
• Applications: Advertising, RLHF, Robotics, …
• Neuro-science and RL

• Normal-form games
• Two-player zero-sum games
• Extensive-form games with imperfect information
• Solving extensive-form games with imperfect information
• Alternatives to Nash equilibrium
• Online learning, regret minimization
• Bayesian games
• Analysis of auctions I (second/first-price auctions as instances of Bayesian games and their equilibria)
• Analysis of auctions II (order statistics, revenue equivalence theorem)
• Introduction to mechanism design (the framework based on Bayesian games, optimal auctions, VCG, revelation principle)
• Coalitional games
• Shapley value
• Weighted voting games
• Summary.

• Introduction and motivation + propositional logic/CNF/BDDs
• SAT - basic encodings, DPLL, resolution
• SAT - CDCL
• SAT - encodings, SMT
• SMT - lazy, congruence closure, theory combination
• Proof assistants
• Intro planning, heuristic search GBFS, A*, Weighted A*
• State space representations PDDL, STRIPS, FDR
• Delete relaxation heuristics hmax, hadd, hff
• Landmarks, saturated cost partitioning, LM-Cut / BDD based symbolic search

LLMs and AGIs, Deep generative models, AI in robotics, Methods of computer vision, Non-smooth non-convex optimization for training deep neural networks