Skip over navigation

Course Offerings

Course Evaluation Results

Course Details

Spring 2016-2017
ELE 571   Graded A-F, P/D/F, Audit

Digital Neurocomputing

Sun-Yuan Kung

Machine learning algorithms allow us to induce rules based on empirical training data. Much of its success lies in an effective representation of objects. Both deep neural networks (DNN) and kernel based machines aim at deriving a much expanded vector space to facilitate larger scale (big) data mining. We shall first cover back-propagation learning algorithms for multi-layer networks, convolution and recurrent neural networks: compare the essential differences between deep and kernel learning paradigms; and explore effective mapping of both DNN and kernel learning models to highly parallel processing architectures.

Sample reading list:
C.M. Bishop, Pattern Recognition and Machine Learning. Berlin: Springer,
S.Y. Kung, Digitial Neural Networks. Prentice Hall, 1993
S.Y. Kung, Kernal Methods and Machine Learning (Cambridge Press, 2014)

Requirements/Grading:
Oral Presentation(s) - 25%
Term Paper(s) - 50%
Problem set(s) - 25%

Other Requirements:
Not Open to Freshmen.

Other information:
The grade will be based on course projects.

Website:  http://www.blackboard.princeton.edu

Schedule/Classroom assignment:

Class numberSectionTimeDaysRoomEnrollmentStatus
42214 S01 9:30 am - 10:50 am M W   Friend Center   111   Enrolled:7 Limit:16