HEAT experiments #625

nkanazawa1989 · 2022-01-25T04:48:08Z

Summary

This PR adds HEAT experiment to the library. This experiment is introduced in the paper https://arxiv.org/abs/2007.02925 which is variant of Hamiltonian tomography with higher sensitivity to small terms. Basically this behaves as pinp-pong experiments.

This is replacement of old PR #460

Details and comments

This PR adds new experiment module qiskit_experiments.library.hamiltonian that is aiming at collecting experiments for low level device characterization. This differs from other characterizations aiming at gate or system metric. This grouping will help users to sort experiments. Of course import path is added to the library so user can still directly import the experiment from there. Note that it is planned to move CrossResonanceHamiltonian to this module in separate PR (and JAZZ is another candidate).

The HEAT experiment is designed as a batch experiment, since this is combination of two ping-pong experiment with different control qubit state. This implementation is more efficient than defining new analysis class that takes two ping-pong curves (i.e. There is no shared parameter between curves. Also we can add as many curves as batch as we want).

Since we can design many HEAT sequences for the different primary term of the generator Hamiltonian and the interrogated error axis, the base class is designed to allow such flexibility. In this PR, most standard ZX-HEAT series are added on top of the base class.

In the separate PR, I plan to add tutorial notebook for 2Q gate characterization with CrossResonanceHamiltonian and HEAT.

- Add hamiltonian tomography experiment module - Add base HEAT framework and analysis to the module - Add API documentation

…riment/heat

eggerdj

Thanks Naoki, this is looking good. Most of my comments are related to docs which need a bit more work.

qiskit_experiments/library/hamiltonian/__init__.py

qiskit_experiments/library/hamiltonian/heat_analysis.py

qiskit_experiments/library/hamiltonian/heat_zx.py

releasenotes/notes/add-heat-experiment-047a73818407e733.yaml

wshanks

I put a few comments in and responded to some of Daniel's. I think this looks really good. I didn't comment much on the documentation because of the many outstanding comments from Daniel.

wshanks · 2022-01-25T20:35:57Z

qiskit_experiments/library/hamiltonian/__init__.py

+# that they have been altered from the originals.
+"""
+===============================================================================================
+Hamiltonian Characterization Experiments (:mod:`qiskit_experiments.library.hamiltonian`)


This is a fairly broad title. Should the contents in cr_hamiltonian.py be moved under qiskit_experiments.library.hamiltonian as well? Also, I put a comment by the reno as well -- I feel like experiments are characterizing a particular schedule more than the underlying Hamiltonian. One could tweak the pulse shape or cancellation tone to adjust observed errors from the target ZX90 operation.

Probably entangler and Entangler characterization? But this could kick out JAZZ experiment from here. I think having static ZZ in the module improves accessibility.

(edit)
Probably we may want to add sizzle too.

qiskit_experiments/library/hamiltonian/__init__.py

wshanks · 2022-01-25T22:09:22Z

qiskit_experiments/library/hamiltonian/heat_base.py

+        The circuit in middle is repeated by ``N`` times to amplify the Hamiltonian
+        coefficients along interrogated axis on the target qubit. The ``prep`` circuit
+        is carefully chosen based on the generator of ``heat`` gate to investigate,
+        and the ``echo`` and ``meas`` circuit depend on the axis of error to amplify.
+        Only target qubit is measured following to the projection by ``meas`` circuit.


That is also the case, though here the meaning was just that only the target qubit is actually measured.

qiskit_experiments/library/hamiltonian/heat_zx.py

releasenotes/notes/add-heat-experiment-047a73818407e733.yaml

test/hamiltonian/test_heat.py

Co-authored-by: Daniel J. Egger <[email protected]>

Co-authored-by: Will Shanks <[email protected]>

…eriments into experiment/heat

…riment/heat

This commit updates the logic of HeatAnalysis constructor according to qiskit-community#633 (now we can instantiate composite analysis with component analyses). Also BatchHeatHelper class is removed and replaced with decorators.

…eriments into experiment/heat

wshanks

Looks good to me! My only remaining quibble is that I see the HEAT experiment as very similar to cr_hamiltonian, so I find it odd they are in separate packages. Perhaps there could be a follow up to move cr_hamiltnonian and possibly give it a more generic name. I suppose it could be used for any gate with a ZX or ZZ Hamiltonian term (just like HEAT it just needs the first term to be Z for the 0 and 1 preparations on the control to make sense).

test/hamiltonian/test_heat.py

eggerdj

This looks mostly good to me. I nit picked the docstrings a bit. My biggest concern on the code front it the decorator in CompositeExperiment. It does not feel right to have a decorator with such a narrow scope.

eggerdj · 2022-02-08T09:11:44Z

qiskit_experiments/framework/composite/composite_experiment.py

@@ -131,3 +132,75 @@ def _postprocess_transpiled_circuits(self, circuits, **run_options):
        for expr in self._experiments:
            if not isinstance(expr, CompositeExperiment):
                expr._postprocess_transpiled_circuits(circuits, **run_options)
+
+
+def sync_transpile_options(


Why is this needed in the first place? I wonder if this should be the default behavior of CompositeExperiment? Or why not have this as an extra function in the CompositeExperiment? I don't see why the decorator framework is needed when we only decorate a single method of a single class.

This aims at making a batch experiment to behave as if a single experiment (because user doesn't need to recognize its implementation to use). This SHOUD not be default behavior (unfortunately it is now) because we should be able to transpile batch (I believe also parallel) experiment with its own configuration.

eggerdj · 2022-02-08T09:15:03Z

qiskit_experiments/framework/composite/composite_experiment.py

+def sync_experiment_options(
+    composite_cls: Type[CompositeExperiment],
+) -> Type[CompositeExperiment]:


Same comment as above. Using decorators here only makes the code harder to follow.

eggerdj · 2022-02-08T09:17:11Z

qiskit_experiments/library/hamiltonian/heat_analysis.py

+
+
+class HeatAnalysis(CompositeAnalysis):
+    r"""A composite error amplification analysis to get unitary error coefficients.


Why not write an example model and make it clear that this has a broader use.

qiskit_experiments/library/hamiltonian/heat_analysis.py

qiskit_experiments/library/hamiltonian/heat_base.py

qiskit_experiments/library/hamiltonian/heat_zx.py

Co-authored-by: Daniel J. Egger <[email protected]>

…riment/heat

Co-authored-by: Daniel J. Egger <[email protected]>

…riment/heat

…eriments into experiment/heat

CLAassistant · 2023-07-18T13:27:22Z

All committers have signed the CLA.

nkanazawa1989 added 5 commits January 20, 2022 01:11

cleanup old PR qiskit-community#460

7b2bfc1

- Add hamiltonian tomography experiment module - Add base HEAT framework and analysis to the module - Add API documentation

write release note

fccf94c

serialization unittest

6e64321

fix projection angle of Z measure

49a7845

Merge branch 'main' of github.com:Qiskit/qiskit-experiments into expe…

b766f56

…riment/heat

nkanazawa1989 requested review from wshanks and eggerdj January 25, 2022 04:48

nkanazawa1989 force-pushed the experiment/heat branch from 9d9cfed to ed50a78 Compare January 25, 2022 05:23

add unittest for running experiments and minor doc fix

d32444f

nkanazawa1989 force-pushed the experiment/heat branch from 49d8096 to d32444f Compare January 25, 2022 06:19

eggerdj suggested changes Jan 25, 2022

View reviewed changes

wshanks reviewed Jan 25, 2022

View reviewed changes

nkanazawa1989 and others added 18 commits January 27, 2022 15:37

Update qiskit_experiments/library/hamiltonian/__init__.py

c0e0da5

Co-authored-by: Daniel J. Egger <[email protected]>

Update qiskit_experiments/library/hamiltonian/__init__.py

423dad4

Co-authored-by: Daniel J. Egger <[email protected]>

Update qiskit_experiments/library/hamiltonian/heat_analysis.py

8adcca2

Co-authored-by: Daniel J. Egger <[email protected]>

Update qiskit_experiments/library/hamiltonian/heat_analysis.py

942e558

Co-authored-by: Daniel J. Egger <[email protected]>

Update qiskit_experiments/library/hamiltonian/heat_base.py

4f70f6e

Co-authored-by: Daniel J. Egger <[email protected]>

Update qiskit_experiments/library/hamiltonian/heat_base.py

9cc247f

Co-authored-by: Daniel J. Egger <[email protected]>

Update qiskit_experiments/library/hamiltonian/heat_base.py

e8a11f7

Co-authored-by: Daniel J. Egger <[email protected]>

Update qiskit_experiments/library/hamiltonian/heat_base.py

00a991a

Co-authored-by: Daniel J. Egger <[email protected]>

Update qiskit_experiments/library/hamiltonian/heat_base.py

7c49da1

Co-authored-by: Daniel J. Egger <[email protected]>

Update qiskit_experiments/library/hamiltonian/heat_zx.py

4a3c645

Co-authored-by: Daniel J. Egger <[email protected]>

Update qiskit_experiments/library/hamiltonian/heat_zx.py

ca871ac

Co-authored-by: Daniel J. Egger <[email protected]>

Update qiskit_experiments/library/hamiltonian/heat_zx.py

8ced3c2

Co-authored-by: Daniel J. Egger <[email protected]>

Update qiskit_experiments/library/hamiltonian/heat_zx.py

52a1439

Co-authored-by: Daniel J. Egger <[email protected]>

Update releasenotes/notes/add-heat-experiment-047a73818407e733.yaml

42f0768

Co-authored-by: Daniel J. Egger <[email protected]>

Merge branch 'main' into experiment/heat

0cec011

document overhaul for ZXHeat class

16fb9ea

document overhaul for HeatElement class

6175036

black

532f855

nkanazawa1989 and others added 12 commits January 31, 2022 13:49

Update qiskit_experiments/library/hamiltonian/heat_zx.py

c4333fb

Co-authored-by: Will Shanks <[email protected]>

Update qiskit_experiments/library/hamiltonian/heat_zx.py

6524858

Co-authored-by: Will Shanks <[email protected]>

Update qiskit_experiments/library/hamiltonian/heat_zx.py

cf47445

Co-authored-by: Will Shanks <[email protected]>

Update qiskit_experiments/library/hamiltonian/heat_zx.py

077d217

Co-authored-by: Will Shanks <[email protected]>

Update qiskit_experiments/library/hamiltonian/heat_base.py

a773faa

Co-authored-by: Will Shanks <[email protected]>

update test tolerance to be more precise

9819c16

Merge branch 'experiment/heat' of github.com:nkanazawa1989/qiskit-exp…

678af55

…eriments into experiment/heat

Update heat_zx.py

1ceff09

Merge branch 'main' of github.com:Qiskit/qiskit-experiments into expe…

e540446

…riment/heat

udpate analysis constructor and remove BatchHeatHelper

e103640

This commit updates the logic of HeatAnalysis constructor according to qiskit-community#633 (now we can instantiate composite analysis with component analyses). Also BatchHeatHelper class is removed and replaced with decorators.

Merge branch 'experiment/heat' of github.com:nkanazawa1989/qiskit-exp…

d918470

…eriments into experiment/heat

fix lint

280cae1

wshanks approved these changes Feb 4, 2022

View reviewed changes

test/hamiltonian/test_heat.py Show resolved Hide resolved

eggerdj suggested changes Feb 8, 2022

View reviewed changes

nkanazawa1989 and others added 13 commits February 9, 2022 07:18

Update qiskit_experiments/library/hamiltonian/heat_base.py

4f7cd81

Co-authored-by: Daniel J. Egger <[email protected]>

Update qiskit_experiments/library/hamiltonian/heat_base.py

5c40d7d

Co-authored-by: Daniel J. Egger <[email protected]>

Update qiskit_experiments/library/hamiltonian/heat_base.py

4882e9e

Co-authored-by: Daniel J. Egger <[email protected]>

Update qiskit_experiments/library/hamiltonian/heat_zx.py

a62cc8f

Co-authored-by: Daniel J. Egger <[email protected]>

Update qiskit_experiments/library/hamiltonian/heat_zx.py

5d4945e

Co-authored-by: Daniel J. Egger <[email protected]>

Update qiskit_experiments/library/hamiltonian/heat_zx.py

c908105

Co-authored-by: Daniel J. Egger <[email protected]>

Merge branch 'main' of github.com:Qiskit/qiskit-experiments into expe…

6c1a556

…riment/heat

Update qiskit_experiments/library/hamiltonian/heat_zx.py

995a009

Co-authored-by: Daniel J. Egger <[email protected]>

Merge branch 'main' of github.com:Qiskit/qiskit-experiments into expe…

caf945a

…riment/heat

remove usage of fitval

c400186

Merge branch 'experiment/heat' of github.com:nkanazawa1989/qiskit-exp…

aba80ca

…eriments into experiment/heat

lint

fec3997

fix bug due to qiskit-community#662

1e21e36

nkanazawa1989 mentioned this pull request Mar 4, 2022

Introducing group to curve analysis #715

Closed

nkanazawa1989 mentioned this pull request Mar 16, 2022

CurveAnalysis refactoring #737

Closed

7 tasks



		class HeatAnalysis(CompositeAnalysis):
		r"""A composite error amplification analysis to get unitary error coefficients.

HEAT experiments #625

Are you sure you want to change the base?

HEAT experiments #625

Uh oh!

Conversation

nkanazawa1989 commented Jan 25, 2022 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Summary

Details and comments

Uh oh!

eggerdj left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

wshanks left a comment

Choose a reason for hiding this comment

Uh oh!

wshanks Jan 25, 2022

Choose a reason for hiding this comment

Uh oh!

nkanazawa1989 Jan 31, 2022 • 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!

Uh oh!

wshanks Jan 25, 2022

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

wshanks left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

eggerdj left a comment

Choose a reason for hiding this comment

Uh oh!

eggerdj Feb 8, 2022

Choose a reason for hiding this comment

Uh oh!

nkanazawa1989 Feb 8, 2022

Choose a reason for hiding this comment

Uh oh!

eggerdj Feb 8, 2022

Choose a reason for hiding this comment

Uh oh!

eggerdj Feb 8, 2022

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

CLAassistant commented Jul 18, 2023 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Uh oh!

nkanazawa1989 commented Jan 25, 2022 •

edited

Loading

nkanazawa1989 Jan 31, 2022 •

edited

Loading

CLAassistant commented Jul 18, 2023 •

edited

Loading