OpenSkillIndex← back to search
claude-codescience

Babysitter rb-benchmarker

Randomized benchmarking skill for gate fidelity characterization

install
source · Clone the upstream repo
git clone https://github.com/a5c-ai/babysitter
Claude Code · Install into ~/.claude/skills/
T=$(mktemp -d) && git clone --depth=1 https://github.com/a5c-ai/babysitter "$T" && mkdir -p ~/.claude/skills && cp -r "$T/library/specializations/domains/science/quantum-computing/skills/rb-benchmarker" ~/.claude/skills/a5c-ai-babysitter-rb-benchmarker && rm -rf "$T"
manifest: library/specializations/domains/science/quantum-computing/skills/rb-benchmarker/SKILL.md
tags
#quantum-computing#error-characterization#fidelity-measurement#rb-protocols#qiskit#cirq
source content

RB Benchmarker

Purpose

Provides expert guidance on randomized benchmarking protocols for characterizing quantum gate fidelities and hardware performance.

Capabilities

  • Standard randomized benchmarking
  • Interleaved randomized benchmarking
  • Simultaneous RB for crosstalk
  • Character benchmarking
  • Cycle benchmarking
  • Fidelity decay fitting
  • SPAM error separation
  • Confidence interval estimation

Usage Guidelines

  1. Protocol Selection: Choose RB variant based on characterization goals
  2. Sequence Generation: Create random Clifford sequences of varying lengths
  3. Execution: Run benchmarking experiments with sufficient statistics
  4. Fitting: Analyze decay curves to extract fidelity parameters
  5. Reporting: Generate comprehensive benchmarking reports

Tools/Libraries

  • Qiskit Experiments
  • Cirq
  • True-Q
  • PyGSTi
  • SciPy