Variational approaches for driven quantum systems
Neural Network variational parametrisation
The potential of Machine Learning (ML) for fundamental research in physics has been recently demonstrated in the context of many-body physics and quantum computing, where exact numerical studies are computationally challenging due to the exponential growth of the dimension of their Hilbert space. The evaluation of variational neural-networks representations of the many-body wavefunction can be achieved with computational performance comparable or better than other variational frameworks such as Montecarlo methods and Tensor networks representations. This recent success of ML applied to quantum science opens the possibility to explore problems in non-equilibrium physics and time-dependent problems. My research focus on devising variational approaches to study the dynamics multimode driven systems, by exploiting the lattice structure of the extended representation of the Hamiltonian.
|