Dr. Pooja Rao is a Senior Quantum Algorithm Engineer at NVIDIA, focused on the intersection of quantum computing, HPC, and AI. Using the open-source CUDA-Q platform, she develops scalable quantum algorithms and applications. Her background in computational fluid dynamics involved designing algorithms and code for large-scale HPC simulations of multiphase turbulent flows.
Presentation Title:
Generative Quantum Eigensolver
Presentation Abstract:
We present the Generative Quantum Eigensolver (GQE), a novel framework that trains classical generative models to build optimal quantum circuits, offering an alternative to standard variational algorithms. Our transformer-based implementation, GPT-QE, outperforms some of the other approaches for ground state search for the electronic structure Hamiltonian. We also demonstrate the method on real quantum hardware.