🐬The PORPOISE Protocol
A web-native, emoji-centric data sharing and commitment standard.
Proof of Reputation Protocol for Outsourced Intuition Sampling Ensembles (PORPOISE) seeks to provide an open protocol for the creation of digital identities that are cryptographically bound to a sequence of concealed predictions about future events. A participant’s predictions can be selectively revealed at a later date or overall prediction accuracy can be attested without disclosing specific event choices.
Predictions are made at verifiable times in regards to specific, publicly posed questions about future events, referred to as “surveys”. The final resolution of the event is also publicly verifiable (e.g. the winner of a local election, or the victor in a sporting event) and committed to the blockchain via prespecified oracles. Lastly, crowd-sourced predictions can effectively be purchased from survey participants through the use of bounties tied to the survey on the blockchain.
The PORPOISE protocol is comprised of the following components:
A human-readable survey posting and data sharing standard, compatible with all modern web-browsers and messenger apps.
A data commitment protocol for survey creators and survey responders.
An identity schema for survey responders to commit to public attributes.
An zero-knowledge circuit that enables a survey responder to prove their PORPOISE score (calculated as the ratio of their correct survey responses to their total responses), with or without exposing their actual choices, based on their historical survey commitments.
A mechanism for tying bounties to surveys with specific requirements that survey responders must meet for eligibility.
A native token for governance and incentive alignment.
Identities are cryptographically bound to a sequence of predictions made at verifiable times about specific, publicly posed questions. Various identity attributes can optionally be verified and committed, including but not limited to:
date of identity creation/age
past predictions and when they were made/updated
association w/ other online identities i.e. social media accounts, known websites, [wallet addresses] etc.
domains of relative expertise
Through the use of zkSNARK technology, an individual's historical prediction performance can be attested to without revealing a participant’s past choices. As its name entails, PORPOISE establishes a set of standards and implementations to aggregate an ensemble of predictions from anonymous participants into a single forward prediction with quantified confidence intervals. Individual predictions can be weighted appropriately based on past performance in an open marketplace where entities can request predictions from participants with provable past performance with a promise of monetary and reputation rewards if their predictions are correct. Conformal prediction offers a solid theoretical framework constructing such confidence intervals.
For our initial deployment, PORPOISE has chosen to integrate with Mina Protocol. Mina ecosystem is particularly advantageous for PORPOISE’s purposes because of O1JS, a tool that simplifies the process of writing cryptographic circuits. Moreover, Protokit in conjunction with the Mina Protocol enables the creation of application-specific sidechains. These sidechains will allow PORPOISE to launch without requiring initial users to hold native PORPOISE or Mina tokens, which is a significant consideration for reducing barriers to entry. This aspect is not only critical from a user accessibility standpoint but also forms a part of our broader strategy to facilitate adoption and engagement through enabling PORPOISE Survey posting across various mainstream social media platforms.
While our initial implementation is Mina-centric, the underlying architecture of our protocol is compatible with any zero-knowledge proof (zkproof) protocol. This design choice ensures that our system remains flexible and can be adapted or extended to other zkproof protocols in the future, maintaining technical robustness and scalability.
Last updated