Quantum Experiment Data Exchange (QEDX) - Building an Interoperability Standard

In this post, we will design the foundations for an interoperability standard for our Quantum Data Management (QDM) Platform. Read more about interoperability in QDM here.


A. Proposed Standard

Quantum Experiment Data Exchange (QEDX)

1. Purpose:

To facilitate the consistent sharing and interpretation of experimental data generated on quantum systems. This includes raw measurement data, metadata, calibration information, and experimental setup descriptions.

2. Scope:

2.1 Data Types:

  • Quantum state preparation and measurement outcomes
  • Quantum circuit execution results
  • Quantum process tomography data
  • Noise characterization data

We are only considering the distinctive QDM platform specific operational data. The consumer/business data assets may be shared through classical “Data Contracts”.

2.2 Metadata:

  • Description of the physical quantum system (number of qubits, architecture, technology)
  • Experimental parameters (control pulse settings, temperatures, etc.)
  • Calibration and error characterization information
  • Processing steps applied to the raw data
  • Provenance: Mechanisms for tracking the origin and history of data

2.3 Format:

  • Base Format: JSON or XML for flexibility, extensibility, and readability.
  • Ontology: Leverages existing ontologies where possible (e.g., QUDT for units of measure) and develops a specialized quantum experiment ontology
  • Schema: A well-defined schema ensures consistency and simplifies parsing of QEDX data.

2.4 Key Principles

  • Open & Non-proprietary: Ensures accessibility and avoids reliance on vendor-specific formats.
  • Extensibility: Allows for representing new types of quantum systems, experiments, and data processing techniques.
  • Completeness: Encourages capturing all relevant information for reproducibility and meaningful interpretation.
  • Machine-Readable: Enables automated data processing and analysis across various tools and platforms.

2.5 Potential Benefits of QEDX

  • Accelerated Research: Easier access to shared, well-structured experimental data fosters faster scientific progress.
  • Reproducibility: Enhances the ability to independently replicate experiments and build upon previous findings.
  • Benchmarking: Facilitates fair comparison of different quantum devices and algorithms on standardized datasets.
  • Collaboration: Enables smoother data exchange between researchers, regardless of their specific experimental setups.


B. Technical Design

1. Core Data Structure

  • Hierarchical Organization: A nested structure to capture relationships between different data elements. Potential top-level sections:

    • experiment: Overall description of the experimental setup and goals.
    • system: Detailing the quantum device used (architecture, qubit technology, connectivity, etc.)
    • calibration: Information on calibration procedures and error characterization.
    • runs: An array of individual experimental runs, each containing:
    • circuit: Description of the executed quantum circuit (if applicable)
    • parameters: Experimental settings (pulse amplitudes, timings, etc.)
    • results: Raw measurement outcomes.
  • Metadata Best Practices:

    • Controlled vocabulary: Leverages existing ontologies where suitable (QUDT, etc.) and extends with a quantum-specific ontology.
    • Timestamps: Include dates and times of experiments.
    • Provenance: Mechanisms to track data lineage (e.g., links to prior datasets used as input)

2. Data Serialization

Below is an example of a JSON based human-readable serialization option. An XML based option may be explored too.

{
  "experiment": "Bell state measurement",
  "system": {
    "type": "superconducting",
    "qubits": 2
  },
  "runs": [
    {
      "circuit": "Bell_circuit.qasm",
      "results": [0, 0, 1, 1]
    }
  ]
}


3. Schema Validation

  • JSON Schema or XSD: Define strict rules for QEDX format adherence.
  • Validation Tools: Ensure data integrity and compliance.
  • Versioning: Mechanism to track schema changes over time for backward compatibility.

4. Noise Characterization Data

Representing noise characterization data within the QEDX format is crucial for making informed decisions about quantum algorithms and error correction strategies.

  • Qubit Characterization:
    • T1 (Relaxation Time): How long a qubit stays in the excited state.
    • T2 (Decoherence Time): Loss of quantum properties over time.
    • Readout errors: Errors in interpreting the state of a qubit.
    • Gate errors: Errors in applying single or multi-qubit quantum gates.
  • Cross-talk: Unwanted interactions between qubits.
  • Environmental Noise: External disturbances (temperature fluctuations, electromagnetic fields).

4.1 Representation in QEDX

Let’s extend the QEDX structure proposed earlier:

4.1.1 Dedicated “calibration” Section:

 "calibration": {
     "noise_characterization": {
         "timestamp": "YYYY-MM-DDTHH:MM:SSZ",
         "methods": ["randomized_benchmarking", ...],
         "results": {
             "qubit_1": {
                 "T1": 50.0,  // Microseconds
                 "T2": 80.0,
                 "readout_error": 0.02, // Probability
                 ...
             },
             "qubit_2": { ... },
             "gate_errors": {
                 "CNOT": {
                     "average_gate_fidelity": 0.95,
                     ...
                 }
             }
             ...
         }
     }
 }


4.1.2 Metadata:

  • methods: Type of characterization techniques used (randomized benchmarking, gate set tomography, etc.).
  • timestamp: Indicate when the calibration data was obtained.

4.1.3 Flexible Results:

  • Structure data by qubit and gate.
  • Include appropriate units and uncertainty estimates.
  • Potentially reference more detailed external data files if needed.

4.2 Evolving Representations

  • Noise Models: QEDX could include ways to represent parameterized noise models derived from characterization data.
  • Dynamic Updates: As noise characteristics fluctuate, mechanisms to update QEDX without full recalibration would be beneficial.

5. Considerations for Additional Functionality

  • Data compression: For large datasets, efficient compression may be necessary.
  • Security: Support for encryption/decryption if handling sensitive data.
  • Data visualization: Recommendations for consistent ways to visually represent quantum experimental data for human interpretation.


C. Final representations

experiment

  • name: A short, descriptive title for the experiment.
  • description: A detailed explanation of the experiment’s goals and procedures.
  • date: Date the experiment was conducted (YYYY-MM-DD format).
  • research_group: The research team or institution responsible.
  • sharing: Level of data visibility (open, restricted, collaborators-only, etc.).
  • access_request_url: Link to a mechanism for requesting access if data is restricted.

system

  • type: Type of quantum system (superconducting, ion trap, photonic, etc.).
  • vendor: Manufacturer of the quantum device.
  • model: Specific model name or identifier.
  • qubits: Number of qubits in the system.
  • topology: Arrangement of qubits and their connectivity.
  • accessible_via: Array listing platforms offering access to this device, with relevant details.

calibration

  • noise_characterization: Section containing noise data.
  • timestamp: When calibration data was collected.
  • methods: Techniques used (e.g., randomized_benchmarking).
  • results: Qubit-specific errors (T1, T2, readout) and gate fidelities.

runs

  • Array of individual experimental runs
  • circuit: Circuit definition with multiple representations if available:
    • qiskit_circuit
    • openqasm
    • cirq_circuit
  • parameters: Execution parameters (shots, simulator settings, etc.).
  • results: Measurement outcomes.
  • data_format: Format of external data, if applicable.
  • external_data_uri: Location of external results data.

metadata

  • quantum_data_platform: Originating platform.
  • provenance: Array for tracking data lineage.
  • ontologies: List of ontologies used in the data.
  • keywords: Terms aiding discoverability.
{
  "experiment": {
    "name": "Bell State Verification",
    "description": "Preparing and measuring a Bell state to assess two-qubit gate fidelity.",
    "date": "2024-05-16",
    "research_group": "Quantum Lab, University X",
    "sharing": "open",
    "access_request_url": "https://qdpexchange.org/request/12345"
  },
  "system": {
    "type": "superconducting",
    "vendor": "Acme Quantum Devices",
    "model": "AcmeQPU-5",
    "qubits": 2,
    "topology": "linear",
    "accessible_via": [
        { "platform": "Acme Cloud Quantum", "region": "US-East" },
        { "platform": "XYZ Quantum Services", "API_endpoint": "https://xyzquantum.api/v1/" },
        { "platform": "IBM Quantum Experience", "backend": "ibmq_ourense" }
    ]
  },
  "calibration": {
    "noise_characterization": {
      "timestamp": "2024-05-15T16:20:00Z",
      "methods": ["randomized_benchmarking"],
      "results": {
          "qubit_0": {
              "T1": 65.0,
              "T2": 90.0,
              "readout_error": 0.015
          },
          "qubit_1": {
              "T1": 58.0,
              "T2": 82.0,
              "readout_error": 0.022
          },
          "gate_errors": {
              "CNOT": {
                  "average_gate_fidelity": 0.965
              }
          }
      }
    }
  },
  "runs": [
    {
      "circuit": {
          "qiskit_circuit": {
              "source": "bell_prep_qiskit.py",
              "version": "0.41.0"
          },
          "openqasm": "bell_prep.qasm",
          "cirq_circuit": {
              "source": "bell_prep_cirq.py",
              "version": "0.16.0"
          }
      },
      "parameters": {
          "shots": 1024,
          "cirq_simulator_id": "noisy"
      },
      "results": [00, 11, 00, 10, ...],
      "data_format": "CSV",
      "external_data_uri": "https://universityx.datarepo/bell_data.csv"
    }
  ],
  "metadata": {
    "quantum_data_platform": "Qiskit QDP",
    "provenance": [
        { "dataset_id": "54321", "source": "Previous calibration run" },
        { "job_id": "63fa8... ", "source": "IBM Quantum job" }
    ],
    "ontologies": ["QEDX-core", "QUDT", "Qiskit-runtime"],
    "keywords": ["Bell state", "fidelity", "superconducting qubits"]
  }
}





    Enjoy Reading This Article?

    Here are some more articles you might like to read next:

  • Introducing ETL-C (Extract, Transform, Load, Contextualize) - a new data processing paradigm
  • (Part 2/3) Rethinking ETLs - How Large Language Models (LLM) can enhance Data Transformation and Integration
  • (Part 1/3) Rethinking ETLs - How Large Language Models (LLM) can enhance Data Transformation and Integration
  • Who Needs Exact Answers Anyway? The Joy of Approximate Big Data
  • Evolutionary Bytes - Harnessing Genetic Algorithms for Smarter Data Platforms (Part 2/2)