How cern-qti/quask works
QuASK is a Python library for constructing quantum kernels, evaluating their properties, and searching for improved kernels. The principal workflow is to define an operation-based ansatz and kernel, execute it through a quantum backend, measure kernel quality, then optionally optimize the kernel. The README also identifies multiple SDK backends and evaluation criteria as major capabilities. The graph represents these stages and the documented capabilities; links without a shown caller or explicit README relationship are omitted rather than inferred.
Main entry points: kernel_evaluator.py Kernel evaluator · ansatz.py Ansatz · core_implementation Backend selection · kernel.py Kernel model
Components 19
Kernel design
- Operations
src/quask/core/operation.py - Ansatz
src/quask/core/ansatz.py - Kernel model
src/quask/core/kernel.py - Kernel factory
src/quask/core/kernel_factory.py
Quantum execution
- Backend selection
src/quask/core_implementation - PennyLane backend
src/quask/core_implementation/pennylane_kernel.py - Qiskit backend
src/quask/core_implementation/qiskit_kernel.py - Braket backend
src/quask/core_implementation/braket_kernel.py - Qibo backend
src/quask/core_implementation/qibo_kernel.py
Kernel analysis
- Kernel evaluator
src/quask/evaluator/kernel_evaluator.py - Alignment measures
src/quask/evaluator - Expressibility measures
src/quask/evaluator - Generalization measures
src/quask/evaluator - Spectral bias
src/quask/evaluator/spectral_bias_evaluator.py - Complexity measures
src/quask/evaluator/ess_model_complexity_evaluator.py
Kernel search
- Optimizer interface
src/quask/optimizer/base_kernel_optimizer.py - Search strategies
src/quask/optimizer - Wide-kernel environment
src/quask/optimizer/wide_kernel_environment.py
Other components
- Library user
Connections 23
- Library user → Ansatz: defines
README.md - Library user → Kernel model: constructs
README.md - Ansatz → Operations: contains
src/quask/core/ansatz.py - Kernel model → Ansatz: uses
src/quask/core/kernel.py - Kernel model → Kernel factory: restores through
src/quask/core/kernel.py - Backend selection → PennyLane backend: offers
src/quask/core_implementation/__init__.py - Backend selection → Qiskit backend: offers
src/quask/core_implementation/__init__.py - Backend selection → Braket backend: offers
src/quask/core_implementation/__init__.py - Backend selection → Qibo backend: offers
src/quask/core_implementation/__init__.py - Qiskit backend → Ansatz: encodes
src/quask/core_implementation/qiskit_kernel.py - PennyLane backend → Ansatz: executes
src/quask/core_implementation/pennylane_kernel.py - Braket backend → PennyLane backend: extends
src/quask/core_implementation/braket_kernel.py - Library user → Kernel evaluator: evaluates
README.md - Kernel evaluator → Alignment measures: includes
src/quask/evaluator/__init__.py - Kernel evaluator → Expressibility measures: includes
src/quask/evaluator/__init__.py - Kernel evaluator → Generalization measures: includes
src/quask/evaluator/__init__.py - Kernel evaluator → Spectral bias: includes
src/quask/evaluator/__init__.py - Kernel evaluator → Complexity measures: includes
src/quask/evaluator/__init__.py - Library user → Search strategies: optimizes
README.md - Search strategies → Optimizer interface: implements
src/quask/optimizer/bayesian_optimizer.py - Search strategies → Wide-kernel environment: uses
src/quask/optimizer/greedy_optimizer.py - Wide-kernel environment → Kernel evaluator: scores
src/quask/optimizer/wide_kernel_environment.py - Wide-kernel environment → Ansatz: updates
src/quask/optimizer/wide_kernel_environment.py