qiskit_noise_learning.experiment_builder.Experiment

class qiskit_noise_learning.experiment_builder.Experiment(*, fidelity_model: LinearMap[Hashable, FidelityIndex] | ModelGateSet | None = None, paths: list[Path] | None = None, instruction_sequences: list[InstructionSequence] | None = None, relations: set[tuple[int, int]] | None = None, shots: int = 20, randomizations: int = 50, randomization_multipliers: list[int] | None = None, validate: bool = True)

GitHub

Bases: object

A learning experiment specification.

An Experiment collects all the data needed to define a noise-learning experiment: a fidelity model, analysis paths, instruction sequences, their relations, and execution parameters (shots and randomizations).

All fields are optional and may be progressively populated via ExperimentBuilderStage instances.

Parameters

fidelity_model – A fidelity model or a model gate set (which is wrapped in an IdentityFidelityModel).
paths – Paths to analyze.
instruction_sequences – Instruction sequences (may include both bound and unbound).
relations – Set of (path_idx, sequence_idx) tuples indicating which paths are traversed by which instruction sequences.
shots – Global number of shots (default 20).
randomizations – Global number of randomizations (default 50).
randomization_multipliers – Per-sequence randomization multiplier (parallel to instruction_sequences).
validate – If True (default), enforce the same validation checks as replace() (co-replacement, length consistency, relations bounds).

init

__init__(*, fidelity_model: LinearMap[Hashable, FidelityIndex] | ModelGateSet | None = None, paths: list[Path] | None = None, instruction_sequences: list[InstructionSequence] | None = None, relations: set[tuple[int, int]] | None = None, shots: int = 20, randomizations: int = 50, randomization_multipliers: list[int] | None = None, validate: bool = True)

Methods

Column 1	Column 2
`__init__`(*[, fidelity_model, paths, ...])
`replace`(*[, validate])	Return a shallow copy with the given fields overridden.

Attributes

Column 1	Column 2
`design_matrix`	The design matrix, lazily computed from the fidelity model and paths.
`fidelity_model`	The fidelity model.
`gate_set`	The model gate set.
`instruction_sequences`	The instruction sequences.
`is_executable`	Whether this experiment has all the information required to be run.
`paths`	The analysis paths.
`randomization_multipliers`	Per-sequence randomization multipliers.
`randomizations`	Global number of randomizations.
`relations`	The set of path and sequence relations.
`shots`	Global number of shots.

fidelity_model

Type: LinearMap[Hashable, FidelityIndex] | None

The fidelity model.

gate_set

Type: ModelGateSet | None

The model gate set.

paths

Type: list[Path] | None

The analysis paths.

instruction_sequences

Type: list[InstructionSequence] | None

The instruction sequences.

relations

Type: set[tuple[int, int]] | None

The set of path and sequence relations.

shots

Type: int

Global number of shots.

randomizations

Type: int

Global number of randomizations.

randomization_multipliers

Type: list[int] | None

Per-sequence randomization multipliers.

design_matrix

Type: IndexedMatrix

The design matrix, lazily computed from the fidelity model and paths.

Raises

ValueError – If fidelity_model or paths is None.

is_executable

Type: bool

Whether this experiment has all the information required to be run.

Requires: instruction_sequences is set, all sequences are bound and complete, and randomization_multipliers is set.

replace

replace(*, validate: bool = True, **kwargs) → Experiment

Return a shallow copy with the given fields overridden.

When validate=True (default), the following checks are enforced:

Co-replacement: instruction_sequences and randomization_multipliers must always both be None or both be non-None. Replacing one without the other is only allowed if it preserves this invariant.
Length consistency: randomization_multipliers must have the same length as instruction_sequences.
Relations bounds: Setting relations requires paths and instruction_sequences to be present, and all indices must be in bounds.
Soft invalidation: Replacing paths or instruction_sequences without providing new relations will set relations to None with a warning.

Parameters

validate – If True, enforce the above checks. If False, fields are set as-is with no validation.
**kwargs – Field names and their new values.

Raises

TypeError – If an unrecognized field name is given.
ValueError – If a validation constraint is violated.

Was this page helpful?

Report a bug, typo, or request content on GitHub.