Can discovering boost learning? - Improving the quality of a machine learning model through discovering hidden structure among data
Abstract
Machine learning can be regarded as a process of identifying certain mapping function from data. The effectiveness of such data-driven framework normally relies heavily on the quality of data. In other words, the performance of a learning model can be seriously affected given missing or noisy data. To address the concern of incomplete input data, the project explores how the performance of a machine learning model (in particular, factorization, ranking, and clustering) can be boosted through incorporating a discovery engine aiming at finding hidden/missing structure or information among data. The project will design a framework that jointly learns a machine learning model and a discovery model with the goal to boost the prediction performance. This project will yield a general framework and concrete case studies about how a learning model can benefit from a discovery model.
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Jul 28, 2017
- Source ID
- FA23861714038
Entities
People
- Shou-De Lin
Organizations
- Air Force Office of Scientific Research
- National Taiwan University
- United States Air Force