Skip to content

Latest commit

 

History

History
58 lines (50 loc) · 1.85 KB

README.MD

File metadata and controls

58 lines (50 loc) · 1.85 KB
Raw

Learning pure states using qWGAN

This document shows how the quantum Wasserstein GAN can be used to learn pure states. An example of the settings in configuration file '''config_pure.py'''.

Preparation

  1. Create a random quantum state as our target state:

We use three direction rotation gates with random angles applied on each qubit to generate the target pure state.

angle = np.random.randint(1, 10, size=[size, 3])
matrix = Identity(size)
for j in range(size):
    row_i_mat = np.matmul(Z_Rotation(size, j, np.pi / angle[j][2], False),np.matmul(Y_Rotation(size, j, np.pi / angle[j][1], False),X_Rotation(size, j, np.pi / angle[j][0], False)))
matrix = np.matmul(row_i_mat, matrix)
real_state = np.matmul(matrix, zero_state)
  1. Define the Generator and Discriminator
  • Create an instance of the generator
gen = Generator(size)
  • Construct and assign the quantum circuit as the generator
gen.set_qcircuit(construct_qcircuit(qc_gen,size))
  • Create an instance of the Discriminator

We define the fixed Hermitian matrices whose linear combination forms the Hermitian

herm = [I, X, Y, Z]
dis = Discriminator(herm, size)

Training

Alternately update the parameters of Generator and Discriminator until the fidelity between the generated quantum state and target state converges to 1.

gen.update_gen(dis,real_state)
dis.update_dis(gen,real_state)

Save the model and plot the figure

After training, the generator and discriminator weights can be saved

save_model(gen, cf.model_gen_path)
save_model(dis, cf.model_gen_path)

and the training loss and fidelity curves can be plotted

plt_fidelity_vs_iter(fidelities, losses, cf)

Output

the fidelity curve and loss curve for learning a 3 qubit pure state Alt text