Research Scientist
CurrentCompleted or am working on a wide range of projects related to machine learning, with a particular emphasis on reinforcement learning. Derived and implemented reinforcement learning approaches to a highly constrained scheduling problem, a blind chess variant, and a question-answer system for augmented image classification. Contributing to a recurrent RL approach for visual saccade. Built a deep network/semantic attribute based framework for zero shot hierarchical image classification. Earlier in my time at APL I tackled a more diverse set of projects including a simulation of noise and spatial correlations in quantum computing circuits, implementations of distillation and transfer learning in deep networks, a hybrid approach for MEG source localization that combines beam forming with change point detection, development and simulation of a quasi-optimal evacuation strategy for crisis situations, and exploration of a free energy-based paradigm for robotic control.
