Nonlinear Dynamics and Deep Learning Researcher,
Aug, 2020 - May, 20254 years 9 months
Dissertation on Differentiable programming with Nonlinear Dynamics
Showed increased expressivity in neural networks through metalearning of activation functions(Published, Patent Pending)
Crafted gradient based optimization framework for reconfigurable nonlinear logic gates(Preprint available, Patent Pending)
Improved neural network convergence by growing neurons in situ during training(Preprint available, Patent Pending)
Controlled noisy chaotic dynamics using self-supervised neural network to configure stochastically integrated complex systems