This dataset contains real-world computational problem instances and their solutions that were generated and validated through blockchain consensus using Proof-of-Useful-Work (PoUW).

Dataset Structure

Schema

Each record contains:

problem_id: Unique identifier for the problem instance
problem_type: "Custom"
problem_data: Custom problem specification (base64-encoded)
- problem_id: Problem identifier
- data: Base64-encoded problem data
solution_data: Custom solution (base64-encoded)
- data: Base64-encoded solution data
problem_complexity: Difficulty weight of the problem
bounty: Token reward for solving (in smallest unit)
time_asymmetry: Ratio of solve time to verify time
space_asymmetry: Ratio of solve memory to verify memory
solve_energy_joules: Energy consumed during solution
verify_energy_joules: Energy consumed during verification
total_energy_joules: Total energy consumed
energy_asymmetry: Ratio of solve energy to verify energy
work_score: Computed work score based on asymmetries
block_height: Blockchain height where solution was accepted
timestamp: Unix timestamp of solution
status: Problem status (Mined, Validated, etc.)
energy_measurement_method: Method used for energy measurement
submission_mode: How problem was submitted (mining, public, private)

Data Fields

All records include comprehensive metrics:

Asymmetry Metrics: Time, space, and energy asymmetries between solving and verification
Energy Measurements: Detailed energy consumption during computation
Work Scores: Blockchain consensus work scores based on computational asymmetries
Blockchain Metadata: Block height, timestamps, miner addresses

Data Splits

This is a continuously growing dataset with new solutions added as they are mined on the blockchain.

Dataset Creation

Source Data

Solutions are generated through the COINjecture Network B blockchain's mining process, where miners solve NP-hard problems to mine blocks.

Energy Measurement

Energy measurements use platform-specific methods:

Linux: RAPL (Running Average Power Limit) interface
macOS: powermetrics
Fallback: CPU TDP-based estimation

Considerations for Using the Data

Bias and Limitations

Problem difficulty varies based on blockchain difficulty adjustment
Energy measurements may use estimation on platforms without hardware monitoring
Solutions represent successfully mined blocks only (failed attempts not recorded)

Additional Information

Dataset Curators

COINjecture Network B - Autonomous blockchain dataset generation

Licensing

MIT License

Citation

@misc{coinjecture_netb_custom,
  author = {COINjecture Network B},
  title = {Custom Problem Type Solutions from Proof-of-Useful-Work Blockchain},
  year = {2025},
  publisher = {Hugging Face},
  url = {https://huggingface.co/datasets/COINjecture/Custom_Solutions}
}

Contact

For issues or questions, please open an issue on the COINjecture GitHub repository.

Generated automatically by COINjecture Network B blockchain nodes

Downloads last month: 8