Dragos D. Margineantu

Robust Machine Learning for decision systems. Learning models that estimate their competence, plan and decide in a reliable manner.

Embedded Machine Learning - designing and training learned models that will be embedded into complex systems with humans and other automated or autonomous components.

Methods for anomaly detection and handling anomalies in high-risk decision systems.

Ensemble learning, Cost-sensitive learning, Active learning.

Inverse Reinforcement Learning. Intent recognition and anomalous action understanding.

Methods for the evaluation and validation of decision systems. The validation and certification of decision systems for deployment.

Knowledge-based learning; integrating domain knowledge into learning algorithms.

Game theoretic decision making and intelligent systems capable of scalable game theoretic decisions.

Geometry and the reasoning required by it. Inference for solving synthetic geometry problems.