authors: Rodrigo A. Vargas-Hernandez, Ricky T. Q. Chen, Kenneth A. Jung, and Paul Brumer
GOAL:
Optimizing quantum non-equilibrium steady state systems using gradient-based methods power by automatic differentiation.
Diagram of the methodology proposed here.
For a V-system,
We maximize the transport of energy to the reaction center (RC) using gradient descent,
All simulations were carried in JAX. We implemented the vector-Jacobian product to differentiate the steady state to incorporate it in the framework of automatic differentiation. For more details about our work we created a Google Collab, where we better exaplain the details of our work.
other systems: We also optimized a three level quantum heat transfer model using gradient based methods. For more details, check our arxiv paper.
Prerequisites: JAX