Artificial Intelligence
and Machine Learning

Probability and learning from evidence · Computation, Turing and early neurons · The birth of AI and early learning systems

AI winters and expert systems · Neural networks, data and computation · Deep learning, generative AI and structured decisions

States, paths, costs and the frontier · Breadth-first and depth-first search · Iterative deepening and lowest-cost-first

Heuristics and greedy search · A* and the conditions behind its guarantees · Branch-and-bound and choosing a strategy

Variables, domains and constraints · Backtracking and variable ordering · Support, propagation and local consistency

Split domains to complete the search · Repair assignments with local search · Escape local minima with annealing

Training, inference and learning signals · Supervised, unsupervised and self-supervised learning · Reinforcement learning and application domains

Features, targets and data tables · Model families, residuals and noise · Training, validation, testing and leakage

Straight lines, residuals and least squares · Fit metrics and held-out evaluation · Polynomial features, overfitting and underfitting

Bias, variance and model selection · Slope tests and confidence intervals · Multiple predictors, confounding and diagnostics

Class probabilities and decision errors · Naive Bayes and its assumptions · Logistic probabilities and thresholds

Likelihood and logistic fitting · Neighbor voting, distance and scaling · Validation, complexity and evaluation

Threshold units and perceptron learning · Linear boundaries and the XOR limitation · Margins, support vectors and soft-margin SVMs

Feature maps and nonlinear boundaries · From feature-space optimization to kernels · Kernel choices, costs and validation

Similarity, centroids and K-means · Choosing the number of groups · Soft membership and expectation–maximization

Hierarchies, linkage and dendrograms · PCA, projection and reconstruction · Scaling, whitening and interpretation