Resonating Perceptron Networks
Department
Mechanical and Aerospace Engineering
Major
Aerospace Engineering
Research Advisor
Wunsch, Donald C.
Advisor's Department
Electrical and Computer Engineering
Funding Source
Missouri S&T OURE Program
Abstract
A novel learning algorithm was developed to increase the speed, stability, and robustness of neural function approximators. A clustering technique is used to identify corresponding regions of the input and output space. These regions and their mappings are then learned independently of all other regions. The resulting behavior is an algorithm which overcomes the problem of catastrophic forgetting while both increasing the speed of learning and minimizing the computational requirements of the network.
Biography
Niklas Melton was born and raised in Kansas City, MO. He will graduate from Missouri S&T this May with a B.S. in Aerospace Engineering. He has been accepted into the Ph.D. program in the Computer Engineering Department beginning in the Fall of 2017. He hopes to continue his fruitful research in neural networks and machine learning far into the future.
Research Category
Engineering
Presentation Type
Poster Presentation
Document Type
Poster
Location
Upper Atrium/Hall
Presentation Date
11 Apr 2017, 1:00 pm - 3:00 pm
Resonating Perceptron Networks
Upper Atrium/Hall
A novel learning algorithm was developed to increase the speed, stability, and robustness of neural function approximators. A clustering technique is used to identify corresponding regions of the input and output space. These regions and their mappings are then learned independently of all other regions. The resulting behavior is an algorithm which overcomes the problem of catastrophic forgetting while both increasing the speed of learning and minimizing the computational requirements of the network.
