INTELLIGENCE-DRIVEN LEARNING

The research objective of this proposal is to investigate implications and applications of the proposed Composite Learning approach that targets the synergy of knowledge-based and data-based methods of machine learning while realizing that synergy as intelligent fusion of distinct components towards achieving state-of-the-art performance with robustness to traditional natural data vulnerabilities and novel adversarial challenges. The anticipated outcome of proposed Composite Learning research will include- (1) Exploring applicability of the proposed approach to calibration machine learning problems. (2) Narrowing the gap between machine learning models in terms of their sensitivity to both natural and malicious data corruption.