Academic year 2015-16
Pattern Recognition
| Degree: | Code: | Type: |
| Bachelor's Degree in Computer Science | 22639 | Optional subject |
| Bachelor's Degree in Telematics Engineering | 22591 | Optional subject |
| Bachelor's Degree in Audiovisual Systems Engineering | 21626 | Optional subject |
| ECTS credits: | 4 | Workload: | 100 hours | Trimester: | 1st |
| Department: | Dept. of Information and Communication Technologies |
| Coordinator: | Xavier Binefa |
| Teaching staff: | Xavier Binefa, Adrià Ruiz |
| Language: | Català, castellà, anglès |
| Timetable: | |
| Building: | Communication campus - Poblenou |
The problem of extracting patterns from data is a success story. It has had results as impressive as the laws of physics or, in more recent, results in areas such as information retrieval, data mining, the character recognition or speech, predicting economic and bioinformatics among others. On all these issues, the issue is to discover regularities that automatically present the data using algorithms based on these regularities and trends, and make decisions such as classifying data into categories. Ultimately -discover- information extracted from the data. The purpose of this workshop is to present general techniques that have proved useful for extracting information. The methods introduced arise from a practical point of view, based on examples.
The course consists of three main activities: lectures, practices, and problems. In the lectures introducing the theoretical concepts and examples of its application. The problems that the students confronted with will be small problems to solve by themselves. Each problem corresponds to one of the concepts introduced in class theory. In practice, problems are more complex so that students have the opportunity to implement a series of concepts learned.
Have taken Probability, Statistics and Stochastic Processes with use. Also a subject Computer Vision may be interesting but not essential.
| General Skills | Specific skills |
|---|---|
|
Instrumentals 1. Cognitive skills 2. Capacity for analysis and synthesis 3. Ability to work with non structured information 4. Organization and planning of personal time Interpersonals 5. Working in group 6. Be competent in presenting the communication Systemic 7. Ability to apply theoretical knowledge in prcactice 8. Relate models with data 9. Ability to generate new ideas |
a) Talking with property in terms of data, models and errors b)Understand and be critical with reading papers on the topics of the course c) Develop empirical habits in the learning of modelization algorithms. d) Knowing how to choose a method of formulating the pros and cons and make contributions to the results. e) To work in areas related to the contents of the subject. f) Implement the processes chosen evaluating the computational cost required.
|
Develop habits empirical learning algorithms modeling
Evaluate each of the three activities of the subject: lectures, practices, and problems. Being:
T: assessment of lectures measured in an exam.
P: assessment of problems: work done in individuals problems.
S: assessment practices. Evaluate the quality of code and memory that handed. There will be three endorsements.
In the continuous assessment, the final note will be obtained as:
Final Note = 0.35T + 0.15*P + 0.5 * S
This formula is applicable only if you get a minimum of four points over ten in the theory exam (T); in the other case you fall the term.
If you need to go to the exam in July, the final note of this examination (with all the lectures) will be:
July Average = 0.6 *JulyExam + 0.4 * Practices. In the case of practices that had been suspended, they can be handed back to practice and assessed again.
T1Introduction
T2 Review and Bayesian Decision Theory
T3 Linear models for classification
T4 Support Vector Classifiers and Kernel methods
T5 Linear Models for Regression
T6 Unsupervised Learning
T7 Latent Variable Models
T8 Combining Classifiers
T9 Advanced topics: Deep Learning
The usual process of learning begins with a theory session in which it is presented some theoretical and practical foundations. The student will complement this activity with a careful reading of his own notes and additional material that the teacher has provided. For example, a 2-hour theory session, properly used, will require additional work outside the classroom 1 hour by the student.
Later made some workout in which the student puts into practice the concepts and techniques presented in the theory session, through the implementation of algorithms for solving small problems or making problems in paper. For the first exercise session will provide solutions, but nothing else. The aim is to consolidate the foundations so that later rise solve more complex problems.
The next step in the learning process is the practice session. It proposed a larger problem requiring a preliminary design and implement a solution that must integrate different concepts and techniques. In the final practice meet all the specific skills that students should acquire in this course.
Kevin P. Murphy. Machine Learning: A probabilistic Perspective. The Mit Press 2012.
S. Theodoridis, K. Koutroumbas An Introduction to Pattern Recognition: A Matlab Approach. Academic Press, 1999.
Ethem Alpaydin, Introduction to Machine Learning, The MIT Press, 2010.
S. Theodoridis, K. Koutroumbas Pattern Recognition, Fourth Edition. Elsevier Academic Press, 1999. http://www.sciencedirect.com/science/book/9781597492720
C. M. Bishop. Pattern Recognition and Machine Learning. Springer 2006.
L. I. Kuncheva Combining Pattern Classifiers, Methods and Algorithms. Wiley Interscience, 2004.
T. Hastie, R. Tibshirani and J. Friedman. The Elements of Statistical Learning: Data mining, Inference, and Prediction. Springer, 2001.
Duda, Hart and Stork,. Pattern Classification, Wiley–Interscience, 2001.
D. Mackay. Information Theory, Inference, and Learning Algorithms. Cambridge University Press, 2003.