HamiltonianGate #4071
HamiltonianGate #4071
Conversation
…ameterExpression.
dongreenberg
requested review from
1ucian0,
ajavadia,
ewinston,
kdk and
mtreinish
as code owners
You can edit |
|
@ajavadia should I? |
| if hasattr(data, 'to_matrix'): | ||
| # If input is Gate subclass or some other class object that has | ||
| # a to_matrix method this will call that method. | ||
| data = data.to_matrix() | ||
| elif hasattr(data, 'to_operator'): |
There was a problem hiding this comment.
Would it be safer to check for types instead of attributes here? The to_matrix is probably safe both ways but to_operator might also be the attribute of another class than BaseOperator.
There was a problem hiding this comment.
I was trying to just replicate the API for UnitaryGate. Do you think it doesn't make sense here?
There was a problem hiding this comment.
To operator lets you initialize with a quantum channel object in the case where that channel is really a unitary channel. This might be a not-needed edge case for this gate.
There was a problem hiding this comment.
Actually - I need something in here to be able to pass in Operators anyway, so is it a problem to leave it as is to handle the broader set of BaseOperators, or would it be better to change it to an isinstance(Operator)?
Co-Authored-By: Julien Gacon <gaconju@gmail.com>
Co-Authored-By: Julien Gacon <gaconju@gmail.com>
Co-Authored-By: Julien Gacon <gaconju@gmail.com>
Co-Authored-By: Kevin Krsulich <kevin@krsulich.net>
kdk
added
the
Changelog: New Feature
There was a problem hiding this comment.
This LGTM, it needs a release note for the new feature but other than that there is the earlier nit that @1ucian0 pointed out in the tests around qobj. Not really a blocker, but would be good to cleanup since you need to iterate again
|
@mtreinish would you mind approving? It says I still need your approval as a codeowner for some reason (even though kdk already approved). |
# Conflicts: # qiskit/extensions/unitary.py
|
I am not a big fan of the name as it makes me think it is not simply a unitary with a Hermitian matrix as the input. This being said i have wasted 20 min thinking of a name and i cant get a better one so im good. |
| --- | ||
| features: | ||
| - | | ||
| New `HamitlotianGate` and `QuantumCircuit.hamltonian()` methods are |
There was a problem hiding this comment.
| New `HamitlotianGate` and `QuantumCircuit.hamltonian()` methods are | |
| New `HamiltonianGate` and `QuantumCircuit.hamiltonian()` methods are |
|
Thanks! Will correct the typo in the changelog |
* Introducing HamiltonianGate, and tests. Tests pass. Allow complex ParameterExpression. * Fix some lint and bugs. * Update qiskit/circuit/parameterexpression.py Co-Authored-By: Julien Gacon <gaconju@gmail.com> * Update qiskit/extensions/hamiltonian_gate.py Co-Authored-By: Julien Gacon <gaconju@gmail.com> * Update test/python/circuit/test_hamiltonian_gate.py Co-Authored-By: Julien Gacon <gaconju@gmail.com> * Update for Julien's changes. * Update for Kevin's changes. * Update for Kevin's changes. * Update test/python/circuit/test_hamiltonian_gate.py Co-Authored-By: Kevin Krsulich <kevin@krsulich.net> * Update test/python/circuit/test_hamiltonian_gate.py Co-Authored-By: Kevin Krsulich <kevin@krsulich.net> * Remove deprecation warning, update for Chris's comments. Tests pass. * Chris's suggestions. * drawer * Make hamiltonian_gate throw an error if QASM is called. * Turn Kevin's changes * Update test/python/circuit/test_hamiltonian_gate.py Co-Authored-By: Kevin Krsulich <kevin@krsulich.net> * Changelog * Remove unnecessary qobj tests per Luciano. * Update for merge. Co-authored-by: Julien Gacon <gaconju@gmail.com> Co-authored-by: Kevin Krsulich <kevin@krsulich.net> Co-authored-by: Kevin Krsulich <kevin.krsulich@ibm.com> Co-authored-by: Luciano Bello <luciano.bello@ibm.com>
Introducing HamiltonianGate, and tests. Tests pass. Also allows complex resolution of ParameterExpression.
Summary
HamiltonianGate takes a hermitian matrix or Operator and time parameter (I wanted to name this
tbut pylint didn't let me), and resolves to a Unitary gate equal to e^iHt. This allows such evolutions to be added to a circuit and the time parameter bound later.Context: I believe this will be a major performance improvement in Aqua simulation, avoiding constructing large evolution circuits for ansatze like UCCSD.