ML-MIPT #389

WandaHou · 2025-05-16T23:42:02Z

This pull request adds code for studying cluster states and measurement-induced entanglement in quantum systems using Cirq. The implementation targets a 6×6 qubit grid with ancilla qubits for readout error mitigation and includes support for circuit construction and data collection.

Key components include:
• Cluster state circuit generation with Hadamard, CZ, and single-qubit rotations
• Basis rotations for measurement in X, Y, Z bases with configurable probe qubits
• Distance-dependent entanglement measurements between probe qubits
• Data collection framework with randomized compiling for error mitigation

The code uses Google’s Cirq framework and is designed to run on quantum processors through Cirq-Google. It builds circuits on a fixed 6×6 qubit grid, with a configurable probe separation distance (d = 3–6) and support for various measurement bases, enabling the study of entanglement properties in cluster states.

The testing framework ensures parameter validation and circuit structure without requiring actual quantum execution, making it suitable for CI/CD pipelines while preserving the ability to run on real quantum hardware.

WandaHou · 2025-05-17T05:54:58Z

Hi, this PR implements the measurement-induced entanglement experiment targeting Google’s superconducting quantum chip. It is part of our collaboration with the Berkeley (Ehud Altman) and Google Quantum AI (Pedram Roushan) teams, under the ML for MIPT project led by Yi-Zhuang You’s group at UCSD.

All tests have passed. Please let me know if any changes are needed. Thank you for your time and review!

eliottrosenberg · 2025-05-17T06:25:30Z

recirq/Cluster_state_MIPT/data_collect.py

+def main() -> None:
+    """Main function to run the data collection."""
+    # Replace these with your specific engine and processor details
+    eng = cg.get_engine('your-engine-id')


This is actually your GCP (Google Cloud) project ID.

eliottrosenberg

@dstrain115 do you want to take a look at this?

@WandaHou are you interested in making an .ipynb notebook in https://github.com/quantumlib/ReCirq/tree/master/docs to explain your experiment?

dstrain115 · 2025-05-17T13:55:42Z

recirq/Cluster_state_MIPT/__init__.py

@@ -0,0 +1,8 @@
+"""Cluster state measurement implementation at MIPT.


Could we change the folder name to be lower case, ie. "cluster_state_mipt" to be consistent with the other folders in recirq?

dstrain115 · 2025-05-17T14:05:22Z

requirements.txt

+black
+torch>=1.9.0


Please add this to recirq/cluster_state_mipt/extra-requirements.txt instead so it is not installed for all modules.
(If we still need this, see above comments)

I have removed the torch installation, and using numpy to save collected data.

Let's remove this from the requirements then.

dstrain115 · 2025-05-17T14:08:02Z

recirq/Cluster_state_MIPT/cluster_state.py

@@ -0,0 +1,119 @@
+import cirq # type: ignore
+import cirq_google as cg # type: ignore
+import numpy as np # type: ignore


Why is this annotated # type: ignore? You should be able to configure mypy to not follow external packages if that's the issue.

Got it, # type: ignore is removed.

dstrain115 · 2025-05-17T14:08:22Z

recirq/Cluster_state_MIPT/cluster_state.py

+import cirq_google as cg # type: ignore
+import numpy as np # type: ignore
+
+def get_two_qubits_6x6(d=6):


Please add a docstring for this function.
Also explain what d is (and maybe change it to a more descriptive name).
If d is the size of the grid, then also change the name of the function, since it is only 6x6 if d=6.

“d” here refers to the spatial separation of the probe qubit pair. In our experiment, it can take values of 3, 4, 5, or 6 in a 6x6 grid. I’ve added a docstring to clarify this.

dstrain115 · 2025-05-17T14:10:29Z

recirq/Cluster_state_MIPT/cluster_state.py

+        [np.array(qubits_matrix).flatten()]+[np.array(anc_pair_set).flatten() for anc_pair_set in anc_pairs]))
+    return qubits_matrix, probe_qubits, anc_pairs, all_qubits
+
+def get_circuit(qubits_matrix, theta, phi, probe_qubits, basis=[0,0], anc_pairs=[]):


Please add docstring.

I have added docstring.

dstrain115 · 2025-05-17T14:14:25Z

recirq/Cluster_state_MIPT/data_collect.py

+
+
+def collect_data(
+    eng: cg.Engine,


nit: recommend engine instead of eng.
Avoid abbreviations, as per:
https://google.github.io/styleguide/pyguide.html#316-naming

I have changed it to engine and also in other places.

dstrain115 · 2025-05-17T14:15:22Z

recirq/Cluster_state_MIPT/data_collect.py

+    theta_range = np.linspace(0, np.pi/2, 11)
+    phi = np.pi*(5/4)
+
+    for d in [5]:


Why is this a for loop if it only has one value?

I changed this to "for d in [3, 4, 5, 6]:" to iterate over different probe qubit distances.

dstrain115 · 2025-05-17T14:16:36Z

recirq/Cluster_state_MIPT/data_collect.py

+    for d in [5]:
+        qubits_matrix, probe_qubits, anc_pairs, all_qubits = get_two_qubits_6x6(d=d)
+        for theta_idx in range(len(theta_range)):
+            for i in range(5):  # iterations


nit: rather than "for i in range(5): # iterations" use "for iterations in range(5):" to be self-documenting.

dstrain115 · 2025-05-17T14:18:29Z

recirq/Cluster_state_MIPT/data_collect.py

+                        measurements = torch.cat([
+                            torch.tensor(r[0].measurements['m']).to(int)
+                            for r in result])
+                        torch.save(


Are we importing all of pytorch just to save a file? Do we really need pytorch to save a matrix to a file?
Could we, instead, use cirq's json serialization instead, for example?

I’ve changed the code to use NumPy for saving the data. Previously, we used PyTorch because the next step in the research involved applying machine learning methods to detect quantum entanglement. However, since we’re now only presenting the circuit construction, I switched to NumPy to eliminate the need for installing PyTorch.

dstrain115 · 2025-05-17T14:19:18Z

recirq/Cluster_state_MIPT/cluster_state.py

@@ -0,0 +1,119 @@
+import cirq # type: ignore


All of these files need to have an Apache copyright header. See other files in ReCirq for an example.

I have added Apache copyright header with year 2025.

WandaHou · 2025-05-17T20:41:17Z

@dstrain115 All the issues should be addressed now—let me know if there’s anything else I should update!

dstrain115

LGTM after the last comments.

dstrain115 · 2025-05-19T14:10:37Z

recirq/cluster_state_mipt/README.md

+
+## Dependencies
+
+- Cirq >= 0.13.0


nit: These dependencies are pretty old (older than the reCirq requirements), so we probably don't need to specify them here, as installing the module via pip will satisfy these.

dstrain115 · 2025-05-19T14:10:49Z

recirq/cluster_state_mipt/README.md

+- Cirq >= 0.13.0
+- Cirq-Google >= 0.13.0
+- NumPy >= 1.19.0
+- PyTorch >= 1.9.0 (for data collection)


Remove if we no longer need this.

dstrain115 · 2025-05-19T14:13:57Z

requirements.txt

+black
+torch>=1.9.0


Let's remove this from the requirements then.

dstrain115

LGTM, thanks for doing the changes!

dstrain115 · 2025-05-19T19:49:11Z

recirq/cluster_state_mipt/__init__.py

@@ -0,0 +1,8 @@
+"""Cluster state measurement implementation at MIPT.


I think even init.py are supposed to have copyright notice.

WandaHou · 2025-05-19T22:40:40Z

It is telling me:

The failing test, test_size_independence_muinf, in the recirq/kpz/experiment_test.py file is caused by the np.isclose assertion for kurtosis.

What do you suggest me to do next step? @dstrain115

dstrain115 · 2025-05-19T22:43:20Z

Could be a flaky test. I re-ran it.

eliottrosenberg · 2025-05-19T22:43:54Z

@WandaHou I think that's my fault for writing a non-deterministic test. I can fix it by passing a fixed seed to the random number generator.

WandaHou · 2025-05-26T21:28:51Z

Hi, just want to make sure, should I do further action to push this into the main branch?

eliottrosenberg · 2025-05-29T03:49:11Z

@WandaHou I know I'm a little late in pointing this out, but you probably also want to update this file to add your work to the table of contents on the website: https://github.com/quantumlib/ReCirq/blob/master/docs/_index_included.yaml.

eliottrosenberg · 2025-05-29T03:50:33Z

Actually, I guess you never made a notebook for the docs folder, so never mind.

WandaHou added 7 commits May 16, 2025 16:27

Create 123.py

2871b6a

Add files via upload

23d8120

Delete recirq/Cluster_state_MIPT/123.py

cba5c4a

Delete recirq/Cluster_state_MIPT/shallow_lattice.pdf

44b161b

Update README.md

869cd3d

Add files via upload

291c679

Add files via upload

b777ad8

WandaHou closed this May 16, 2025

WandaHou reopened this May 16, 2025

WandaHou added 19 commits May 16, 2025 17:00

Update __init__.py

7ca06f3

Add files via upload

e28dfef

Update pythonpackage.yml

bc754da

Delete recirq/Cluster_state_MIPT directory

2668f70

Create README.md

2534e6a

Add files via upload

048d673

Delete recirq/Cluster_state_MIP directory

cf6e10a

Create README.md

78cf0fd

Add files via upload

c65a089

Add files via upload

c78a475

Add files via upload

efd82d4

Delete recirq/Cluster_state_MIPT directory

6d6eef4

Create README.md

bc66462

Add files via upload

34f1d72

Update pythonpackage.yml

e7aa4c7

Delete recirq/Cluster_state_MIPT directory

1a84881

Create README.md

1bec2d4

Add files via upload

019a89f

Update requirements.txt

4738b4e

Update data_collect.py

a54e4bc

eliottrosenberg reviewed May 17, 2025

View reviewed changes

eliottrosenberg approved these changes May 17, 2025

View reviewed changes

Update data_collect.py

258f687

dstrain115 reviewed May 17, 2025

View reviewed changes

WandaHou added 3 commits May 17, 2025 13:19

Delete recirq/Cluster_state_MIPT directory

ddabe49

Create README.md

0d9a50d

Add files via upload

a6aef3f

dstrain115 reviewed May 19, 2025

View reviewed changes

WandaHou added 2 commits May 19, 2025 12:46

Update README.md

7e255bc

Update requirements.txt

8482e45

dstrain115 approved these changes May 19, 2025

View reviewed changes

WandaHou added 2 commits May 19, 2025 13:07

Update __init__.py

46652bf

Merge branch 'master' into ML-MIPT

e17cb0a

Merge branch 'quantumlib:master' into ML-MIPT

12e40a7

dstrain115 merged commit 00b51f6 into quantumlib:master May 26, 2025
5 checks passed

eliottrosenberg mentioned this pull request May 29, 2025

make test deterministic #391

Merged

		@@ -0,0 +1,8 @@
		"""Cluster state measurement implementation at MIPT.

		black
		torch>=1.9.0

		black
		torch>=1.9.0

ML-MIPT #389

ML-MIPT #389

Uh oh!

Conversation

WandaHou commented May 16, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

WandaHou commented May 17, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

eliottrosenberg May 17, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

eliottrosenberg left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

WandaHou commented May 17, 2025

Uh oh!

dstrain115 left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

dstrain115 left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

WandaHou commented May 19, 2025

Uh oh!

dstrain115 commented May 19, 2025

Uh oh!

eliottrosenberg commented May 19, 2025

WandaHou commented May 16, 2025 •

edited

Loading

WandaHou commented May 17, 2025 •

edited

Loading

eliottrosenberg May 17, 2025 •

edited

Loading