TL;DR: Quantum Hamiltonian Descent is a recently proposed algorithm that uses quantum tunneling to escape local minima in non-convex optimization landscapes. Unlike classical gradient descent, which can get
Quantum Hamiltonian Descent is a recently proposed algorithm that uses quantum tunneling to escape local minima in non-convex optimization landscapes. Unlike classical gradient descent, which can get stuck in local minima, quantum tunneling allows the algorithm to pass through energy barriers. This could be particularly useful for training deep neural networks and molecular geometry optimization.
Complexity
Potential speedup over gradient descent for non-convex optimization
Application
Non-convex optimization, training neural networks, molecular geometry optimization