Fermi-Hubbard spin-charge separation results

{ }

View on QuantumAI Run in Google Colab View source on GitHub Download notebook

This notebook presents the experimental data which was collected on Google Rainbow processor for the Fermi-Hubbard spin-charge separation experiment.

try:
    import recirq
except ImportError:
    print("Installing ReCirq...")
    !pip install git+https://github.com/quantumlib/recirq --quiet
    print("Installed ReCirq!")

import glob
from tqdm.notebook import tqdm

from recirq.fermi_hubbard import (
    InstanceBundle,
    apply_rescalings_to_bundles,
    find_bundles_rescalings,
    load_experiment,
    plot_quantity
)
from recirq.fermi_hubbard.publication import (
    parasitic_cphase_compensation,
    fetch_publication_data
)

# Hide numpy warnings
import warnings
warnings.filterwarnings('ignore')

Get the data

In order to run this notebook, the data sets gaussians_1u1d.zip, trapping_2u2d.zip and trapping_3u3d.zip need to be downloaded and extracted from https://doi.org/10.5061/dryad.crjdfn32v. The function fetch_publication_data is a utility to do this.

data_dir = "fermi_hubbard_data"
fetch_publication_data(base_dir=data_dir)

Downloading gaussians_1u1d_nofloquet...
Successfully downloaded.

Downloading gaussians_1u1d...
Successfully downloaded.

Downloading trapping_2u2d...
Successfully downloaded.

Downloading trapping_3u3d...
Successfully downloaded.

Noninteracting Gaussians

# Load results and create a bundle with extracted quantities.
gaussians_1u1d_files = glob.glob(f'{data_dir}/gaussians_1u1d/0.0/*.json')
gaussians_bundle = InstanceBundle(
    experiments=[load_experiment(file) for file in gaussians_1u1d_files],
    steps=range(65),
    rescale_steps=range(65))

# Simulate the exact numerical results that are used as a reference.
with tqdm(range(len(gaussians_bundle.steps))) as progress:
    def post_run(_1, _2):
        progress.update()
    gaussians_bundle.cache_exact_numerics(post_run_func=post_run)

0%|          | 0/65 [00:00<?, ?it/s]

plot_quantity(gaussians_bundle, 'post_selection', show_std_dev=True);

png

plot_quantity(gaussians_bundle, 'scaling', show_std_error=True);

png

# The data for this quantity can be viewed after double-clicking this cell output.
plot_quantity(gaussians_bundle, 'up_down_density', show_std_error=True);

png

plot_quantity(gaussians_bundle, 'up_down_position_average', show_std_error=True);

png

plot_quantity(gaussians_bundle, 'up_down_position_average_dt', show_std_error=True);

png

Trapping Potential N=4

# Load results and create a bundles with extracted quantities for each
# interaction strength.
trapping_2u2d_files = [
    glob.glob(f'{data_dir}/trapping_2u2d/{u}/*.json')
    for u in [0.0, 1.0, 1.5, 2.0, 2.5, 3.0, 3.5, 4.0, 4.5, 5.0]]
trapping_2u2d_bundles = [InstanceBundle(
    experiments=[load_experiment(file) for file in files],
    numerics_transform=parasitic_cphase_compensation(0.138),
    steps=range(11),
    rescale_steps=range(11)) for files in trapping_2u2d_files]

# Simulate the exact numerical results that are used as a reference.
total_steps = sum(len(bundle.steps) for bundle in trapping_2u2d_bundles)
with tqdm(range(total_steps)) as progress:
    def post_run(_1, _2):
        progress.update()
    for bundle in trapping_2u2d_bundles:
        bundle.cache_exact_numerics(post_run_func=post_run)

0%|          | 0/110 [00:00<?, ?it/s]

# Use shared rescaling values among compatible problem instances.
apply_rescalings_to_bundles(find_bundles_rescalings(trapping_2u2d_bundles))

plot_quantity(trapping_2u2d_bundles, 'post_selection', show_std_dev=True);

png

plot_quantity(trapping_2u2d_bundles, 'scaling', show_std_error=True);

png

plot_quantity(trapping_2u2d_bundles, 'charge_spin_density', show_std_error=True);

png

plot_quantity(trapping_2u2d_bundles, 'charge_spin_spreading', show_std_error=True);

png

plot_quantity(trapping_2u2d_bundles, 'charge_spin_spreading_dt', show_std_error=True);

png

Trapping Potential N=6

# Load results and create a bundles with extracted quantities for each 
# interaction strength.
trapping_3u3d_files = [
    glob.glob(f'{data_dir}/trapping_3u3d/{u}/*.json')
    for u in [0.0, 1.0, 1.5, 2.0, 2.5, 3.0, 3.5, 4.0, 4.5, 5.0]]
trapping_3u3d_bundles = [InstanceBundle(
    experiments=[load_experiment(file) for file in files],
    numerics_transform=parasitic_cphase_compensation(0.138),
    steps=range(11),
    rescale_steps=range(11)) for files in trapping_3u3d_files]

# Simulate the exact numerical results that are used as a reference.
total_steps = sum(len(bundle.steps) for bundle in trapping_3u3d_bundles)
with tqdm(range(total_steps)) as progress:
    def post_run(_1, _2):
        progress.update()
    for bundle in trapping_3u3d_bundles:
        bundle.cache_exact_numerics(post_run_func=post_run)

0%|          | 0/110 [00:00<?, ?it/s]

# Use shared rescaling values among compatible problem instances.
apply_rescalings_to_bundles(find_bundles_rescalings(trapping_3u3d_bundles))

plot_quantity(trapping_3u3d_bundles, 'post_selection', show_std_dev=True);

png

plot_quantity(trapping_3u3d_bundles, 'scaling', show_std_error=True);

png

plot_quantity(trapping_3u3d_bundles, 'charge_spin_density', show_std_error=True);

png

plot_quantity(trapping_3u3d_bundles, 'charge_spin_spreading', show_std_error=True);

png

plot_quantity(trapping_3u3d_bundles, 'charge_spin_spreading_dt', show_std_error=True);

png