PennyLaneAI / PennyLaneAI/catalyst
Contextual decomposition strategy to improve differentiability
Nobody has claimed this yet.
- Dominant language
- Python
- Stars
- 234
- Forks
- 84
- Avg merge
- 2d 15h
- Merged PRs (30d)
- 66
Description
Issue description
-
Expected behavior: QubitUnitary for gates with more than 2 qubits is not differentiable using Enzyme. This means that programs that transform gates to QubitUnitary cannot be differentiated. However, it is possible to differentiate these programs if instead of transforming gates to QubitUnitary they are decomposed to other gates.
-
Actual behavior: The decomposition strategy right now is indeed to decompose as much as possible without changing them to QubitUnitary, but once matrix gates are used, some gates will be changed to QubitUnitary.
Contributor guide
First steps
- Read the whole issue, then the project's contributing guide.
- Comment on the issue to say you are picking it up — it saves two people doing the same work.
- Fork the repository and make your change on a branch.
- Open a pull request that references the issue number.
Research direction
No files or tests are named. Start by locating the decomposition strategy and the handling of QubitUnitary for matrix gates, then determine how contextual decomposition should preserve differentiability with Enzyme. Done means programs that would otherwise become multi-qubit QubitUnitary operations can be differentiated after decomposition.
Written by the indexing model from the issue text.
Assessment
- Domain
- compilers
- Issue type
- Bug
- Difficulty
- 5/5
- Estimated time
- Over a week
- Activity status
- Stale
- Clarity
- Needs clarification
- Newbie friendliness
- 20/100