Pattern Recognition Using Bio-inspired Classification
In this project, the students will explore the use of a bio-inspired algorithm for pattern recognition called Hierarchical Temporal Memory (HTM). HTM is based on Jeff Hawkins' (founder of Palm Computing) book on On Intelligence, in which he describes his theory of how the human neocortex works. He believed that the neocortex works on common algorithm where sensory inputs such as vision, hearing, touch, language, behavior and almost everything else the neocortex are manifestations of a single cortical algorithm. (More information on HTM is available at http://numenta.com)
The project aims to:
Test the performance of HTM on some benchmark data for pattern recognition problems. Through this project, the students will gain better understanding of the problem of pattern recognition, the considerations in developing good solutions to it, and in addition be exposed to an emerging research direction of bio-inspired pattern recognition.
The challenge of the project:
HTM is a very recent technology and a strong mathematical foundation is necessary to understand and analyse its behavior.
Interest in algorithm development, proficient programming skills in MATLAB, JAVA and Python (optional), strong mathematical and analytical skills