MLQMB school quizz

What is the computation complexity of generating a spin configuration using an RNN wave function in terms of the system size N ?

Using a positive RNN wave function implies that

• All the parameters are positive 

• All the amplitudes are positive 

• All the phases are positive 

• The variational energy is positive 

An RNN wave function correspond to

• An energy-based model

• An auto-regressive model

• A generative adversarial model

What makes RNN wave-functions very attractive for the study of quantum many-body systems?

Which kinds of dynamics can be treated using time-dependent variational Monte Carlo?

• Fully time-dependent Hamiltonians

• Only quenches with constant Hamiltonians

Which of these quantities are conserved by time-dependent variational Monte Carlo (as introduced in the lectures) ?

• Energy (for a time-independent Hamiltonian)

• Integrals of motion (i.e., observables commuting with the Hamiltonian)

• Norm of the quantum state

• Norm of the variational parameter vector

Which properties of the biological brain does neuromorphic hardware aim to capture?

•  Slow & lazy

•  Fast & energy-efficient

• Small & parallelized

• Noisy & chaotic

Which quantum states can straightforwardly be reconstructed with a standard RBM?

• Ground states of stoquastic Hamiltonians

• Highly excited states

•  Time-evolved states

• All strongly entangled states

Which neural network architecture provides the best wavefunction ansatz?

• Restricted Boltzmann Machine

•  Recurrent Neural Network

• Spiking Neural Network

• Convolutional Neural Network

• Number of visible neurons

• Number of hidden neurons

• Number of Gibbs sampling steps per training iteration

•  Number of variational parameters