Babes-Bolyai University of Cluj-Napoca
Faculty of Mathematics and Computer Science
Study Cycle: Master
SUBJECT
| MII1002 | Machine Learning |
| Teaching Staff in Charge |
Assoc.Prof. CZIBULA Gabriela, Ph.D., gabis cs.ubbcluj.ro |
| Aims |
|
1. To provide an introduction to the basic principles, techniques, and applications of Machine Learning.
2. To cover the principles, design and implementation of learning programs which improve their performance on some set of tasks by experience. 3. To offer a broad understanding of machine learning algorithms and their use in data-driven knowledge discovery and program synthesis. 4. To offer an understanding of the current state of the art in machine learning in order to conduct original research in machine learning. |
| Content |
|
1. Introduction in Machine Learning. Statistical Foundations
- Issues in Machine Learning - Designing a learning system - Event space and Probability function - Elementary Information Theory 2. Decision Tree learning - Decision tree representation - ID3 learning algorithm - Statistical measures in decision tree learning: entropy, information gain - Issues in DT learning 3. Artificial Neural Networks - Neural Network representations - Appropriate problems for Neural Network Learning - Perceptrons - Multilayer Networks and the Backpropagation algorithm - Advanced topics in Artificial Neural Networks 4. Support Vector machines - Main idea - Linear SVMs - Non-linear SVMs - Applications 5. Instance based learning - k-Nearest Neighbor learning - Locally weighted regression - Radial basis functions - Case based reasoning 6. Bayesian learning - Bayes theorem - Naive Bayes Classifier - Bayesian Belief Networks - EM algorithm - Examples 7. Learning set of rules - Sequential covering algorithms - Learning Rule Sets - Learning Sets of First Order Rules (FOIL) - Induction as Inverted Deduction - Inductive Logic Programming 8. Unsupervised Learning - Cluster analysis - Self organizing maps - Invatare Hebbiana - Applications 9. Reinforcement Learning - The reinforcement learning task - Markov Decision Processes - Q-learning - Temporal Difference learning - Applications |
| References |
|
1. Mitchell, T., Machine Learning, McGraw Hill, 1997
2. Russell, J.S, Norvig, P., Artificial Intelligence- A Modern Approach, Prentice- Hall, Inc., New Jersey, 1995 3. Gabriela Czibula, Sisteme inteligente. Instruire automata, Ed. Risoprint, Cluj-Napoca, 2008 4. Manning, C., Schutze, H., Foundations of Statistical NLP, MIT Press, 2002 5. Cristiani, N., Support Vector and Kernel Machines, BIOwulf Technologies, 2001 6. Nillson, N., Introduction to Machine Learning, Stanford University, 1996 7. Sutton, R.S., Barto, A.G., Reinforcement learning, The MIT Press Cambridge, Massachusetts, London, England, 1998 8. Şerban, G., Pop, H.F., Tehnici de Inteligenţă Artificială. Abordări bazate pe Agenţi Inteligenţi, Ed. Mediamira, Cluj-Napoca, 2004 9. Şerban, G., Sisteme multiagent în Inteligenţa Artificială Distribuită. Arhitecturi şi aplicaţii. Editura RisoPrint, Cluj-Napoca, 2006 |
| Assessment |
|
The activity ends with a written final exam (grade E). During the semester, the students will have to prepare a theoretical report (grade R) and a practical project that will have to illustrate a learning task (grade P). The students activity during the semester will be also considered (grade A). The final grade is the weighted mean of the four grades mentioned above. The final grade = 40%E + 25%R + 25%P + 10%A. Students who demonstrate excellent research performance by developing the project to publication will get an extra score of 10% from the final grade. Successful passing of the exam is conditioned by the final grade that has to be at least 5. More about evaluation can be found at http://www.cs.ubbcluj.ro/~gabis/Req_ML.htm. |
| Links: | Syllabus for all subjects Romanian version for this subject Rtf format for this subject |