Babes-Bolyai University of Cluj-Napoca
Faculty of Mathematics and Computer Science
Study Cycle: Graduate
SUBJECT
| MID0037 | Advanced Methods of Machine Learning |
| Teaching Staff in Charge |
Assoc.Prof. CSATO Lehel, Ph.D., csatol cs.ubbcluj.ro |
| Aims |
|
The aim of the course is to familiarise students with modern machine learning methods. At the end of the module the student will be able to analyse the outputs of different algorithms and to decide on which one to use when faced with real applications. |
| Content |
|
Machine learning is the application of algorithms to various types of data. The emphasis is on a careful examination of the type of the data. It is commonplace that more data means more information. However, to extract the information, usually one has to include prior knowledge into the algorithm. Prior knowledge is needed to select the model we want to use. The employed algorithm depends on the size and type of available observations – called data. If there are huge data-sets available, then one must use statistical methods, for a few data a lot more emphasis is to be put on the model. The algorithms depend also on the type of the problem we want to solve. In this respect we prefer an algorithm that is developed for classification when classification is to be solved and we would use specific algorithms for discrete data and other set for continuous observations. The course is aimed as a practical introduction to various models from machine learning.
Thematic overview of the lectures: •Component analysis (weeks 1-4) ref. [2,4,8,9]: • Basics of statistical modeling – matrices and eigenvalues, • Principal components used in signal de-noising, • Independent components used in blind source separation of signals. • Bayesian modeling (weeks 5-7) ref. [1,2,3,4]: • Introduction to various estimation methods, • Hierarchical model specification and parameter estimation, • Bayesian modeling: obtaining the posterior and the predictive distributions. • Hidden Markov models (HMMs) weeks (8-11) ref. [6,7]: • HMM definitions, • Estimating the latent states, • Applications of HMMs (1) speech recognition, • Applications of HMMs (2) gene segmentation, • Gaussian Process models (weeks 12-14) ref. [4,5]: • Joint Gaussians, kernel functions, Functional Bayesian models, • Approximations: using the KL-projection, sparse approximations, • Applications of Gaussian process inference. |
| References |
|
[1]. Bishop C.M (2006) Pattern Recognition and Machine Learning, Springer Verlag.
[2]. Russell S, Norvig P (2003) Artificial Intelligence: A Modern Approach (Second Edition), Prentice Hall. [3]. Mitchell T (1997) Machine Learning, McGraw Hill. [4]. Bernardo J.M, Smith A.F.M (2000) Bayesian Theory, John Wiley & Sons. [5]. MacKay D.J.C (2003) Information Theory, Inference and Learning Algorithms, Cambridge University Press, HTTP: http://wol.ra.phy.cam.ac.uk/mackay/itila/book.html. [6]. Rasmussen C.E, Williams C.K.I (2006) Gaussian Processes for Machine Learning, The MIT Press. [7]. Rabiner L.R, Juang, B.H (1986) An introduction to Hidden Markov models, IEEE ASSP Magazine, pp: 4-15. [8]. Durbin R, Eddy S.R, Krogh A, Mitchison G (1999) Biological Sequence Analysis: Probabilistic Models of Proteins and Nucleic Acids. Cambridge University Press. [9]. Hyvärinen A, Karhunen J, Oja E (2001) Independent Component Analysis, Wiley-Interscience. [10].Barto A. (2002): Statistical Pattern Recognition, John Wiley & Sons. |
| Assessment |
|
The exam will consist of (40%) a presentation of a topic chosen in the first 6 weeks of the semester, (20%) the solution of the laboratory examples, (40%) oral examination based on the topics of the lectures and the seminars presented by the students. |
| Links: | Syllabus for all subjects Romanian version for this subject Rtf format for this subject |