This Course
Purpose : Pattern recognition is a study to mathematically simulate human's capability of recognition and cognition. Applications of pattern recognition including face recognition, fingerprint recognition, speech recognition and object detection and categorization have become state-of-the-art technologies in the cloud and IoT eras. This course will give you a mathematical foundation for these applications. However, more in-depth materials will be given to deep neural networks, which is a state-of-the-art approach in pattern recognition and machine learning.
English teaching: This course will proceed with English. In addition to English lectures given by the teacher, reports of home works and one project have to be done with English. Moreover, interactions among class students such as group discussions and activities will also be with English.
Evaluation: Evaluation of student's performance is based on a multitude of metrics, including reading reports, oral presentation, programming results, group collaboration, and peer review. Programming skills including Matlab/C/C++ is necessary to practice and implement the deep learning method. Some topics in the course will be presented by students. Interactive forms of in-classroom activities will be planned in the course. A project will be assigned with paper reading, program coding, oral presentation and report writing. Project is done by team work. Some presentations and reports are evaluated by peer review.
English teaching: This course will proceed with English. In addition to English lectures given by the teacher, reports of home works and one project have to be done with English. Moreover, interactions among class students such as group discussions and activities will also be with English.
Evaluation: Evaluation of student's performance is based on a multitude of metrics, including reading reports, oral presentation, programming results, group collaboration, and peer review. Programming skills including Matlab/C/C++ is necessary to practice and implement the deep learning method. Some topics in the course will be presented by students. Interactive forms of in-classroom activities will be planned in the course. A project will be assigned with paper reading, program coding, oral presentation and report writing. Project is done by team work. Some presentations and reports are evaluated by peer review.
Reference Books and Materials
- Neural Networks and Deep Learning, by Michael Nielsen, 2015. (Free online book)
- Deep Learning, MIT Press, in preparation, Y. Bengio, I. Goodfellow, A. Courville, 2015. (Free PDF)
- L. Deng, D. Yu. “Deep learning: methods and applications.” Foundations and Trends in Signal Processing, NOW Publishers, 7.3–4, 197-387, 2014. (Free PDF)
- Yoshua Bengio, "Learning Deep Architectures for AI," Foundations and Trends in Machine Learning, 2(1), pp.1-127, 2009. (Free PDF)
- Deep Belief Nets in C++ and CUDA C, by Timothy Masters, 2015. (Book information)
- Pattern Recognition, 4th, S. Theodoridis & K. Koutroumbas, Academic Press, 2009. (Book's web site)
- Introduction to Pattern Recognition - A Matlab Approach," S. Theodoridis, A. Pikrakis, K. Koutroumbas, D. Cavouras, Academic Press, 2010.
Announcement, Course Materials, Homeworks.
Grading
- Reading assignments : 40%
- Paper study and presentation : 20%
- Group discussion : 15%
- Project : 20%
- Presence : 5%
Office Hour: Wednesday 13:30-15:30
Requirements
- English: listening, reading, writing and speaking; in-class English discussion is required.
- Programming skill: Matlab/C/C++/Python.
- Background courses: digital image processing, computer vision.