Incentive Mechanism

Subnet 1 incentive and operation

The core idea behind the subnet is to share problem-solving code openly to enhance competition results, round by round. Open-sourcing these solutions widens impact and helps tackle real-world challenges.

In the Matrix Compression competition — the first event on the subnet — the winning solution will not only directly enhance Subnet 9’s output but may also contribute to optimising data centres and cloud infrastructure.

General Operations

• Solution Submission Miners submit their solutions to the competition to the submission endpoint API.

• Execution and Metrics Each solution runs inside an isolated secure sandbox, generating its evaluation metrics.

• Evaluation The subnet validator reviews these metrics and assigns a score to each submitted solution.

• Scoring and Selection Validators retrieve the scores from the subnet orchestrator to determine the winning miner.

• Rewards The miner with the highest score receives the full reward. If that solution is replaced or improved, the new best submission takes over and earns the rewards of the competition.

Incentive Challenges

1. How does SN1 prevent successful solution code from being stolen and resubmitted for reward?

To ensure miners are fairly rewarded for their best solutions, the submitted code remains hidden for a set period. Once evaluation is complete and rewards are distributed, the code is made public for others to explore.

The winner’s code is revealed with a delay. The duration of the hidden code period depends on the competition. Logs are published only after the completion of a competition round.

This approach guarantees creators to fully exercise the submission rewards, while still allowing others to learn from and improve existing solutions — tapping into the power of community innovation.

2. How are miners driven toward continuous improvement?

The emission burning mechanism was created to push successful miners toward improving their code. Once a top-performing solution is submitted, its associated emissions gradually start to burn. The longer the solution remains unchallenged, the higher the burn rate becomes — motivating continuous innovation and competition.

3. What will be the baseline and success metrics of the Matrix Compression competition?

The initial baseline for the Matrix Compression challenge will be the Lempel–Ziv–Markov-Chain Algorithm (LZMA). LZMA, recognisable in due to its frequent usage in 7-zip. It combines dictionary and range encoding, saving bits on repeated data and predictable byte sequences. Apex adopts the implementation from the standard python library.

4. How are code submissions validated?

The Code Executor is a part of the subnet where miner code is ran and evaluated. Evaluation metrics depend on the competition. In case of the Matrix Compression competition such metrics are:

• Compression ratio - the max level of compression provided by solution

• Compression speed - the time required to run a compression and decompression processes

After solutions run in the Code Executor, metrics are converted into score in accordance with the given competition's scoring mechanism.