Proof-of-Weights (PoWg)
Last updated
Last updated
Proof-of-Weights (PoWg) is a decentralized scoring framework used to evaluate and rank inference nodes in the Hubic network. Validators compute performance metrics for AI agents or inference nodes and must prove — via zero-knowledge proofs — that scoring was executed correctly.
This zk-verified scoring ensures a reliable reputation system, which is essential for both network health and RWA-linked models that generate income based on performance.
🔧 How It Works:
Validator retrieves performance data for a model or inference node.
Scoring function is executed locally, encapsulated in a zk-circuit.
A zk-proof of execution is generated and submitted on-chain.
Scores and proofs are recorded and validated via Ethereum smart contracts.
🔐 Security Guarantees:
Prevents score manipulation or duplication across validators.
Validates that scoring was computed honestly using on-chain verifiable circuits.
Publishes score hashes and proofs to Ethereum for auditability and downstream automation.
🔁 Flow:
📊 Use Cases:
Node Reputation: Reliable scoring ensures high-performing nodes are prioritized for critical inference tasks.
Model Quality Tracking: Models with better usage metrics and inference success rates can receive more traffic and fees.
RWA Revenue Optimization: RWA-linked models or agents can be scored transparently, with zk-proven logic determining payouts to token holders based on usage and performance quality.
🌍 RWA Relevance:
In the RWA context, PoWg is the accountability engine. It provides cryptographic proof that a model or AI agent is performing as expected — forming the basis for royalty distribution, dynamic pricing or trust-weighted access. High-score agents can command premium rates, attract capital flows and participate in enterprise or institutional deployments.