Algorithms that defy the gravity of learning curve
Abstract
The grantee has developed a new theory of mass in relation to machine learning and developed a very efficient/effective way of estimating mass distribution of data given only a small sample using an ensemble technique, and applied this to construct multiple learners for different tasks: classification, clustering, anomaly detection and information retrieval, each of which exhibited better performance than the state-of-the-art algorithms. This raised a fundamental question: why this defies the conventional wisdom "the gravity of learning curve", (i.e., more data is expected to produce better performing models).
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Mar 23, 2016
- Source ID
- FA23861514009
Entities
People
- Kai Ming Ting
Organizations
- Air Force Office of Scientific Research
- Federation University Australia
- United States Air Force