• Representation and interpretation of the concept Machine Learning (ML), goals of ML, types of learning, supervised and unsupervised learning, bias, designing of learning system
• Ordering of the version space, version space search, search from general to specific, search from specific to general, over-specification, over generation, candidate elimination algorithm
• Representation and use of logical conjunctions, inducing logical conjunctions, EGS algorithm, heuristic induction, HGS algorithm
• Induction of the disjunction normal form, non-incremental learning using separate and conquer principle (NSC), direct generation of classification rules (AQ11)
• Nonincremental generation of clasification decision trees (ID3, C4.5), incremental induction of decision trees (ID5R), trees pruning, nonincremental generation of regretion decision trees (CART), linear versus unlinear data and task
• Induction of decision lists, direction with the aid of exceptions (NEX), ordered and non-ordered set of rules (CN2), numeric characteristics of decision rules
• Representation and use of threshold concepts, criteria tables, heuristic induction of criteria tables (HCT), linear and spherical threshold units, iterative weight perturbation (IWP)
• Representation and use of SVM (Support Vector Machines), solving linear and nonlinear problem, using Kernel funcion
• Instance based learning, representation and use of prototypes, instance averaging, induction of competitive disjunctions (NCD, ICD)
• Probabilistic representation and classification, induction of naive Bayesian classifiers
• Extensional versus intentional concept representation, lazy learning, derivation of the gradient descent rule, locally weighted regression
• Acquiring search control knowledge, reinforcement learning, bucket brigade, Q - learning
• Unsupervised learning, iterative distance-based clustering, K-means, conceptual clustering, CLUSTER/2 algorithm, hierarchical clustering, COBWEB algorithm, density-based clustering, DBSCAN, probabilistic clustering
• Learnig using set of methods, final decision by application of voting, random forests (RF), stacking, bagging, boosting, federated learning
• Principles of neural networks, deep networks, reccurent neural networks, LSTM, GRU, attention mechanism, Transformer, BERT family, T5 family, GPT
• Computational learning theory, probably learning and approximately correct hypothesis, error of a hypothesis, PAC learn-ability, sample complexity for finite and infinite hypothesis spaces, Vapnik-Chervonenkis dimension
• Basic principles of machine learning, cognitive algorithm design, improvement of the classification algorithm results (bagging and boosting methods)

Copyright © kM