title |
---|
Yuma Consensus |
The Bittensor API is designed to enable subnet owners to write their own incentive mechanisms. These incentive mechanisms allow the subnet validators to express their own subjective preferences about what the network should learn. Such an approach:
- Facilitates the economic market in which producers (subnet miners) are constantly driven to make their knowledge output more useful in terms of speed, intelligence and diversity.
- And also decentralizes Bittensor's governance across multiple diverse stakeholders, ensuring that no single group has full control over what is learned.
A subnet validator in a subnet expresses their perspective about how performant subnet miners in the subnet are, through a set of weights
Such weights
For example, the below code prints the netuid
of 1
. You can print such netuid
.
import bittensor as bt
subnet = bt.metagraph( netuid = 1, lite = False)
print ('weights', subnet.W )
The Yuma Consensus algorithm translates the weight matrix
However, radical divergence from consensus view points is dangerous, especially if bad actor validators manipulate incentives in ways that benefits themselves, for example, lying about the value produced by miners.
To avoid this scenario Bittensor uses a mechanism called Yuma Consensus. The Yuma Consensus rewards subnet validators with dividends for producing miner-value evaluations that are in agreement with the subjective evaluations produced by other subnet validators, weighted by stake.
The below example code prints the values of S
, subnet validator stake, and W
, subnet validator weights for a subnet with the netuid
of 1
:
import bittensor as bt
subnet = bt.metagraph( netuid = 1, lite = False)
print ('subnet 1 validator stake', subnet.S )
print ('subnet 1 validator weights', subnet.W )