آموزش هوش مصنوعی: یادگیری ماشین و مبانی تئوری (Artificial Intelligence - Learning & Theory) - آخرین آپدیت

Concept learning

• 1-1 Brief Introduction to Machine Learning, Learning from Example

• 1-2 Hypotheses ,Relation between Instance Space and Hypotheses

• 1-3 The Find-S Algorithm

• 1-4 Version Space and The List-Then Eliminate Algorithm

• 1-5 The Candidate Elimination Algorithm

• 1-6 Biased and Unbiased Hypothesis Space, Futility of Bias-Free Learning

Computational Learning Theory

• 2-1 Introduction to Computational Learning Theory, Setting of Sample Complexity

• 2-2 Setting 3, PAC Learnable

• 2-3 Exhausting the Version Space: Definition, Theorem ,Proof and some examples

• 2-4 Shatter, Dichotomy, VC dimension

• 2-5 Some examples and discussion about VC dimension

• 2-6 Upper and Lower Bounds on Sample Complexity with VC dimension, The Mistake Bound for Algorithms

• 2-7 Optimal Mistake Bound

• 2-8 The Weighted-Majority Algorithm and its Bound

Classification

• 3-1 Decision Trees and its Hypothesis Space

• 3-2 Learning Decision Tree, Information

• 3-3 Generalization and Overfitting, Kai Square Pruning,Rule Post-Pruning

• 3-4 Model Evaluation: Metrics for Performance Evaluation, Methods for Model Comparison

• 3-5 Ensemble: Embedding, Bagging and Boosting

• 3-6 Support Vector Machine: Optimization, Soft Margins, and Kernel Trick

Neural Network and Deep learning

• 4-1 Introduction to Neural Network

• 4-2 Single-Layer Network and Perceptron Learning Rule

• 4-3 Multi-Layer Perceptron, Back Propagation Learning, Decline of ANN

• 4-4 Cascade Correlation Neural Networks, Deep or Shallow Structure

• 4-5 Deep Learning: Convolutional Neural Networks

• 4-6 LeNet 5, Dropout, ReLU and the Variants, Maxout, Residual Net

• 4-7 Recurrent Networks, Long Short-Term Memory (LSTM), Neural Turing Machine, Memory-Augmented Neural Networks (MANN)

• 4-8 Autoencoder: Denoising Autoencoder, Stacked Autoencoder and Variational Autoencoder

• 4-9 Generative Adversarial Net (GAN), AE+GAN and Its Applications

Reinforcement learning

• 5-1 Basic Term of Markov Decision Process, Sequential Decisions, Policy, Utility

• 5-2 Derivation of Bellman Equation

• 5-3 Value iteration and Policy Iteration

• 5-4 Q-learning, Learning Policy,Simple GLIE Scheme

• 5-5 Temporal Difference Algorithm,Generalization

• 5-6 Deep Q-learning and Improvement

• 5-7 Deep Policy Network, Partially Observable MDP,Summary