QSBench/QSBench-Depolarizing-Demo-v1.0.0
收藏资源简介:
--- license: cc-by-nc-4.0 task_categories: - tabular-regression - feature-extraction language: - en tags: - qiskit - quantum-circuits - synthetic-dataset - benchmark - expectation-values - quantum-computing - qml-benchmark - quantum dataset - qml dataset - quantum benchmark - noisy quantum data - depolarizing noise - error mitigation - noise robustness pretty_name: QSBench Depolarizing Demo v1.0.0 – Noisy Quantum Dataset (Depolarizing Noise, n=6) size_categories: - 1K<n<10K ---  🌐 [Website](https://qsbench.github.io) | 🤗 [Dataset](https://huggingface.co/datasets/QSBench/QSBench-Depolarizing-v1.0.0-demo) | 🛠️ [GitHub](https://github.com/QSBench/QSBench-Depolarizing-v1.0.0-demo) | 🚀 [Interactive Demo](https://huggingface.co/QSBench/spaces) # QSBench Depolarizing Demo v1.0.0 **Quantum Machine Learning dataset for noise robustness and error prediction.** Includes paired ideal and noisy expectation values under depolarizing noise. Keywords: quantum dataset, noisy quantum circuits, depolarizing noise, QML benchmark, expectation value prediction. **5000 synthetic quantum circuits with depolarizing noise** — demo subset of the QSBench Noise Pack. Designed for researchers and engineers working on noise-aware quantum ML, robustness analysis, and error mitigation. ### Why this dataset? Real quantum hardware is noisy. Depolarizing noise is one of the most widely studied noise models in quantum computing. This dataset allows you to: - Compare **ideal vs noisy outputs** - Train models that **predict or correct noise effects** - Benchmark **robustness of ML models** - Study **error distributions in quantum circuits** ### Use Cases - Noise robustness benchmarking - Error mitigation research - Predicting noisy expectation values - Learning error correction models - Comparing clean vs noisy quantum outputs ### Dataset Overview - **Samples**: 5000 - **Qubits**: 6 - **Depth**: 4 - **Circuit Families**: Mixed (HEA, RealAmplitudes, QFT, Efficient SU(2), Random) - **Entanglement**: Full - **Noise**: Depolarizing (p = 0.02–0.03 range) - **Observables**: Z, X, Y in mixed mode (global + per-qubit) - **Shots**: 512 - **Splits**: Train / Validation / Test — deterministic hash-based ### What's Inside Each Sample Each sample in the Parquet files contains: - Raw and transpiled QASM representations - Circuit adjacency matrix - Gate statistics (CX, H, RX, RY, RZ, etc.) - Structural metrics: Gate entropy + Meyer-Wallach entanglement - **Ideal expectation values** - **Noisy expectation values** (after depolarizing noise) - **Explicit error targets**: `error_<label> = ideal - noisy` - Circuit metadata and generation parameters - Deterministic split label ### Key Learning Signals For every observable, the dataset provides: `ideal_expval_*`, `noisy_expval_*`, `error_*`, `sign_ideal_*`, `sign_noisy_*`. This enables complex regression and classification tasks for noise modeling. ## QSBench-Depolarizing: Quantum Error Prediction **You don't need a PhD in Quantum Physics to use this dataset.** Think of this as a classic **Predictive Maintenance** or **Regression** problem. We have a machine (the quantum computer) that makes errors. We give you the blueprints of the tasks it ran, and the magnitude of the errors it made. ### The ML Mission: Supervised Regression Your goal is to predict the `error` without running expensive physics simulations. Can you train a Gradient Boosting model (XGBoost/LightGBM) or a Neural Network to predict the output error based purely on the circuit's structural features? ### Dataset Anatomy (Features & Targets) Use the structural features as `X`, and the errors as `y`. | Group | Column Name | What is it for ML? | | :--- | :--- | :--- | | **Features (X)** | `depth`, `gate_entropy`, `cx_count` | The structural complexity of the task. | | **Features (X)** | `noise_prob`, `shots` | The environmental conditions (error probability and sampling rate). | | **Target (y)** | `error_Z_global` | **The Main Target.** The continuous error value you want to predict. | | **Target (y)** | `sign_ideal_Z`, `sign_noisy_Z` | **For Classification.** Did the noise flip the final answer? (Binary target). | ### Quick Start Idea Build a robust XGBoost regressor using `total_gates`, `depth`, and `noise_prob` to predict `error_Z_global`. What is your Mean Absolute Error (MAE)? ### Load the Dataset The dataset is stored in Parquet format inside the `data/shards/` folder. You can load it directly using the Hugging Face `datasets` library: ```python from datasets import load_dataset # Load the depolarizing demo dataset dataset = load_dataset("QSBench/QSBench-Depolarizing-v1.0.0-demo", split="train") # Inspect the first sample with noise data print(dataset[0]) ``` ### Repository Structure The dataset is stored in the `main` branch and contains only the data files to ensure the Dataset Viewer works correctly: ```text QSBench-Depolarizing-v1.0.0-demo/ ├── README.md # This file └── data/ # Parquet shards (main data) └── shards/ └── *.parquet └── *.csv ``` All metadata files (coverage.json, schema.json, meta.json, etc.) are located in a separate branch called `meta`. 👉 browse [meta branch](https://huggingface.co/datasets/QSBench/QSBench-Depolarizing-v1.0.0-demo/tree/metadata) ### Related QSBench Datasets - QSBench Lite (20k samples, n=4) - QSBench Core (75k samples, n=8) - Depolarizing Noise Pack (150k samples) - Amplitude Damping Pack (150k samples) - Transpilation Hardware Pack (200k samples) ### Part of the QSBench Family This is a small public **demo version**. Full-scale datasets (20k–150k+ samples), specialized noisy versions, and custom hardware packs are available. [Repository](https://github.com/QSBench/QSBench-Depolarizing-v1.0.0-demo) | [Website & Full Catalog](https://qsbench.github.io) ### Notes This dataset is fully synthetic and generated using quantum circuit simulation. No real-world or personal data is included. **License:** CC BY-NC 4.0 (Personal & Research Use) **Questions or custom requests?** Visit our [website](https://qsbench.github.io) or open an issue on [GitHub](https://github.com/QSBench/QSBench-Depolarizing-v1.0.0-demo), or inspect the generation pipeline in the **QSBench Generator** [repository](https://github.com/QSBench/QSBench-Generator). ### Support QSBench You can support the project directly on this Giveth page: **[https://giveth.io/project/qsbench](https://giveth.io/project/qsbench)** Your donations help us generate larger datasets, cover GPU costs, and continue developing new realistic noise models. --- *Generated with QSBench Generator v5.0.2*
--- license: cc-by-nc-4.0 任务类别: - 表格回归(tabular-regression) - 特征提取(feature-extraction) 语言: - 英语(en) 标签: - Qiskit(qiskit) - 量子电路(quantum-circuits) - 合成数据集(synthetic-dataset) - 基准测试(benchmark) - 期望值(expectation-values) - 量子计算(quantum-computing) - 量子机器学习基准测试(qml-benchmark) - 量子数据集(quantum dataset) - 量子机器学习数据集(qml dataset) - 量子基准测试(quantum benchmark) - 带噪声量子数据(noisy quantum data) - 去极化噪声(depolarizing noise) - 误差缓解(error mitigation) - 噪声鲁棒性(noise robustness) 显示名称:QSBench去极化演示版v1.0.0——带噪声量子数据集(去极化噪声,n=6) 样本量范围: - 1K<n<10K ---  🌐 [官网](https://qsbench.github.io) | 🤗 [数据集](https://huggingface.co/datasets/QSBench/QSBench-Depolarizing-v1.0.0-demo) | 🛠️ [GitHub仓库](https://github.com/QSBench/QSBench-Depolarizing-v1.0.0-demo) | 🚀 [交互式演示](https://huggingface.co/QSBench/spaces) # QSBench去极化演示版v1.0.0 **用于噪声鲁棒性与误差预测的量子机器学习数据集。包含去极化噪声下的理想与带噪声期望值配对数据。** 关键词:量子数据集、带噪声量子电路、去极化噪声、QML基准测试、期望值预测。 包含5000个带去极化噪声的合成量子电路——为QSBench噪声包的演示子集。 专为从事噪声感知量子机器学习、鲁棒性分析与误差缓解研究的科研人员与工程师设计。 ## 为何选择本数据集? 真实量子硬件存在噪声,而去极化噪声是量子计算中研究最广泛的噪声模型之一。本数据集可用于: - 对比**理想与带噪声输出** - 训练用于**预测或修正噪声影响**的模型 - 基准测试**机器学习模型的鲁棒性** - 研究**量子电路中的误差分布** ## 应用场景 - 噪声鲁棒性基准测试 - 误差缓解研究 - 带噪声期望值预测 - 学习误差校正模型 - 对比干净与带噪声量子输出 ## 数据集概览 - **样本数**:5000 - **量子比特数**:6 - **电路深度**:4 - **电路家族**:混合类型(HEA、RealAmplitudes、量子傅里叶变换(QFT)、高效SU(2)、随机电路) - **纠缠方式**:全纠缠 - **噪声类型**:去极化噪声(概率范围0.02–0.03) - **可观测量**:混合模式下的Z、X、Y(全局+单量子比特) - **采样次数**:512 - **数据划分**:训练集/验证集/测试集——基于哈希的确定性划分 ## 每个样本包含内容 Parquet文件中的每个样本包含: - 原始与转换后的QASM表示 - 电路邻接矩阵 - 门统计信息(CX、H、RX、RY、RZ等) - 结构指标:门熵+Meyer-Wallach纠缠度 - **理想期望值** - **带噪声期望值**(经过去极化噪声后) - **显式误差目标**:`error_<label> = 理想值 - 带噪声值` - 电路元数据与生成参数 - 确定性划分标签 ## 关键学习信号 对于每个可观测量,数据集提供`ideal_expval_*`、`noisy_expval_*`、`error_*`、`sign_ideal_*`、`sign_noisy_*`,可用于构建复杂的回归与分类任务以进行噪声建模。 --- # QSBench-Depolarizing:量子误差预测 **无需量子物理博士学位即可使用本数据集。可将其视为经典的**预测性维护**或**回归**问题。我们拥有一台会产生误差的机器(量子计算机),为您提供其运行任务的蓝图以及产生的误差幅度。** ## 机器学习任务:监督回归 您的目标是无需运行昂贵的物理模拟即可预测`error`。您能否仅基于电路的结构特征,训练梯度提升模型(XGBoost/LightGBM)或神经网络来预测输出误差? ## 数据集结构(特征与目标) 以结构特征作为`X`,以误差作为`y`。 | 特征组 | 列名 | 机器学习用途 | | :--- | :--- | :--- | | **特征(X)** | `depth`、`gate_entropy`、`cx_count` | 任务的结构复杂度 | | **特征(X)** | `noise_prob`、`shots` | 环境条件(误差概率与采样率) | | **目标(y)** | `error_Z_global` | **主目标。** 待预测的连续误差值 | | **目标(y)** | `sign_ideal_Z`、`sign_noisy_Z` | **分类任务。** 噪声是否翻转了最终结果?(二分类目标) | ## 快速入门思路 使用`total_gates`、`depth`和`noise_prob`构建鲁棒的XGBoost回归器以预测`error_Z_global`。您的平均绝对误差(MAE)是多少? ## 加载数据集 数据集以Parquet格式存储在`data/shards/`文件夹中。您可以直接使用Hugging Face的`datasets`库加载: python from datasets import load_dataset # 加载去极化演示数据集 dataset = load_dataset("QSBench/QSBench-Depolarizing-v1.0.0-demo", split="train") # 查看首个带噪声数据的样本 print(dataset[0]) ## 仓库结构 数据集存储在`main`分支中,仅包含数据文件以确保数据集查看器正常工作: text QSBench-Depolarizing-v1.0.0-demo/ ├── README.md # 本文件 └── data/ # Parquet分片(主数据) └── shards/ └── *.parquet └── *.csv 所有元数据文件(coverage.json、schema.json、meta.json等)位于名为`meta`的独立分支中。👉 浏览[meta分支](https://huggingface.co/datasets/QSBench/QSBench-Depolarizing-v1.0.0-demo/tree/metadata) ## 相关QSBench数据集 - QSBench Lite(20000个样本,n=4) - QSBench Core(75000个样本,n=8) - 去极化噪声包(150000个样本) - 振幅阻尼包(150000个样本) - 编译硬件包(200000个样本) ## 属于QSBench家族 这是一个小型公开**演示版本**。完整规模的数据集(20000–150000+个样本)、专用噪声版本以及自定义硬件包均可获取。 [GitHub仓库](https://github.com/QSBench/QSBench-Depolarizing-v1.0.0-demo) | [官网与完整目录](https://qsbench.github.io) ## 注意事项 本数据集为完全合成数据集,通过量子电路仿真生成,未包含任何真实世界或个人数据。 **许可证**:CC BY-NC 4.0(个人与研究用途) **疑问或定制请求?** 请访问我们的[官网](https://qsbench.github.io)或在[GitHub](https://github.com/QSBench/QSBench-Depolarizing-v1.0.0-demo)上提交Issue,或在QSBench生成器[仓库](https://github.com/QSBench/QSBench-Generator)中查看生成流水线。 ## 支持QSBench项目 您可以通过此Giveth页面直接支持本项目: **[https://giveth.io/project/qsbench](https://giveth.io/project/qsbench)** 您的捐赠将帮助我们生成更大规模的数据集、覆盖GPU成本,并持续开发新的逼真噪声模型。 --- *使用QSBench Generator v5.0.2生成*



