DeCartography
DeCartography is a Relational Computation Oracle capable of leveraging Web3 crowdsourcing to provide credible Neutrality social graphs. Dapps utilizing this oracle can make use of quantified data representing cooperation and social distance as a social graph.
One specific use case of DeCartography is in determining the distribution and priority of funding for public goods. The social identity layer of DeCartography will be used for this purpose. This will be employed in collaboration with our initial beta experiment partner, Gitcoin, to alleviate “collusion” in Quadratic Funding — a situation where socially close agents collaborate to ensure a specific project receives more grant money.
What sets DeCartography apart is its unique way of generating social graphs based on the Schelling Point mechanism, by harnessing Web3 crowdsourcing. While there are existing social graph providers that also “analyze given data to create a social graph and provide the results to other apps,” DeCartography positions Crowd-workers as validators of a decentralized oracle. This oracle assigns certain tasks and equilibrates based on the concept of the Schelling Point, forming a computation oracle that can be considered a Proof-of-(Human)-Work.
This mechanism differs from conventional decentralized oracle validators in that where traditional validator nodes are computers (a computational/mathematical approach), DeCartography employs a complex system determined by the intuition and common sense of Crowd-workers.
The former requires predefined measurement criteria due to its analytical nature. In contrast, DeCartography uses crowdsourcing and peer prediction methods to bypass the need for specific analysis metrics for generating social graphs. This approach is based on the hypothesis that by “If you decide on an indicator, it will be hacked” (Perverse incentive) and can be avoided.
- As a Developer: If you have data you want to analyze, send that data along with the required fee to the DeCartography platform. After a certain period, the analyzed data will be returned. Currently, in beta, we are mainly analyzing data from projects that have reached out to us.
- As a Crowd-worker: By solving tasks given by DeCartography as a Crowd-worker, you can earn cryptocurrency without needing KYC. Sign in to DeCartography using your wallet, complete tasks, wait for the specified period, and the reward will automatically be transferred.
Cartography and Credible Neutrality
As the name “DeCartography” (Decentralized + Cartography) suggests, DeCartography is a governance technology for creating maps. In essence, it is a Dapp that leverages crowdsourcing to generate a social graph quantifying contributor collaboration.
Zooming out, we need to acknowledge that there won’t be a one-size-fits-all model to measure and apply social graphs to Plurality. So instead, let’s build a world where a Plurality of social institutions can query soulbound journals based on custom algorithms to assess and distribute rights and responsibilities to their people. The process might become a vital foundation for plural mechanism design across social organizations and mechanisms, including democracies, markets, the data economy, the commons, and identity. Research and development of identity and policy modules for Plurality might become a whole new business model for Gitcoin. For example, we might work with diverse ecosystems on grant rounds and, beyond quadratic funding, provide Plurality infrastructure for democratic, self-organizing communities more generally.
Notably, for example, as the infrastructure scales, it might help solve even some of the most significant technical issues of our time: e.g., Sybil-resistance, digital authentication, and more. Check out this piece on intersectional social data to learn more about some of these possible applications.
https://gov.gitcoin.co/t/how-soulbound-tokens-can-make-gitcoin-grants-more-pluralistic/10077
As Leon Erich mentioned above, DeCartography is expected to be used as the infrastructure for various Dapps as an identity and policy module for “Plurality”.
The term ‘Cartography’ originates from the Greek words χάρτες (map) and γραφειν (write). DeCartography is a project that allows us to represent the state of objective reality (on-chain activity of each wallet) in a reliable way, by adding a level of abstract concepts (plotting to a matrix, dimension reduction, clustering).
To achieve this, it’s essential to ensure credible neutrality from the mechanism layer. But what does “Credible Neutrality” mean in this system?
Vitalik Buterin describes a good mechanism in his piece “Credible Neutrality As A Guiding Principle” as a tool that takes inputs from multiple people, uses these inputs to determine the participants’ values, and makes some decisions that people care about. In a well-functioning mechanism, the decisions made by the mechanism are efficient (meaning the decision considers the participants’ preferences to achieve the best possible outcome) and compatible with incentives, meaning people have incentives to participate “honestly”.
For the DeCartography team to not manipulate the outcome of the “cartography” to their advantage, it’s necessary to establish credible neutrality. This includes:
- Not embedding specific individuals or outcomes into the mechanism.
- Publicly verifiable execution through open source.
- Keeping things simple.
- Avoid frequent changes.
To maintain these rules, DeCartography is built in a public space, where anyone can refer to the following resources:
- All source code is publicly available on GitHub and git.decartography
- All change histories are published globally, and all updates are announced.
Of course, the social graph data as an output of DeCartography is generated from the input of Crowd-workers as participants in the mechanism. However, maintaining this balance while the core team builds with momentum, particularly in the early stages, is expected to be challenging. To overcome this, we plan to build while gathering public opinions and criticisms, iterating through cycles called “Rounds”.
Additionally, to ensure credible neutrality towards Crowd-workers, DeCartography adopts the Peer Prediction Method as mentioned in Vitalik’s Credible Neutrality article. The Peer Prediction Method refers to mechanisms for extracting information from human agents when the information obtained from each agent cannot be directly verified (there is any Grand Truth). These mechanisms are designed to have a game-theoretic equilibrium where everyone reveals their information honestly.
By implementing these measures, DeCartography aims to create an environment where reliable mapping technology and credible neutrality coexist, setting the stage for a more democratically organized future.
Adopting Schelling Point-Based Proof-of-(Human)-Work in a Decentralized Computation Oracle
The hypothesis and mechanism of DeCartography are entirely dependent on the concept of Schelling Point (Focal Point).
Consider a coordination game where you and another prisoner are placed in separate rooms and a guard gives you both a piece of paper with several numbers. If both of you select the same number, you will be set free; otherwise, in game theory terms with little concern for human rights, you will be locked in solitary confinement for life. The numbers are as follows:
14237, 59049, 76241, 81259, 90215, 100000, 132156, 157604
Which number do you pick?
Theoretically, all these numbers are arbitrary, and the chances of getting out of prison by randomly picking the same number are 1/8. But in reality, most people would choose 100000, making the odds significantly higher. Why 100000? Because each prisoner believes that the number 100000 is “special”, and each prisoner believes that the other prisoner also believes that the number 100000 is “special”. Therefore, each prisoner believes that the other is likely to pick 100000, and so they choose 100000 themselves. Clearly, this is a chain of logic that recursively goes on indefinitely and ultimately has no “backing” other than itself.
The tasks that DeCartography issues to Crowd-workers are fundamentally based on this principle. Specifically, it goes as follows:
- After displaying 9 wallet addresses and the NFTs they hold, Crowd-workers select 3 that appear “similar”.
- After repeating this task 10 times (1 session), they send their answers, signed via a smart contract.
- After the Round ends, all the Crowd-workers’ answers are aggregated, and the correlation is calculated.
- Voting power is calculated based on the Humanity Score obtained from the Gitcoin Passport and the amount staked before the task.
- Based on the correlation and voting power, the reward percentage for each agent is calculated, and the payment is made from the Round’s Budget.
Why does this seemingly cunning mechanism work? Essentially, for the same reason that the prisoner’s dilemma worked well above. We assume that the common sense of what appears to be “similar” is a Schelling point. People want to provide their intuition because they expect others to provide the same answer, and the protocol encourages it.
The reason for using an oracle to cluster a list of addresses and output a social graph, even though it takes tens of times longer and costs much more than using a computer, is that we believe there is value in analyzing based on external data from Crowd-workers. The codes running on Ethereum are fundamentally all public, so if you had to write somewhere the code to generate the social graph… For instance, “Cluster people who have Crypto Punks as ‘NFT’”, “Cluster people who have more than 100 NFTs as ‘NFT’”, “Cluster the pool of Uniswap as…”. If you wrote code to generate the social graph like this, they can be seen from the outside and avoided.
Therefore, we hypothesize that it would be beneficial to have an oracle driven by people’s intuition.
Issues and Limitations
The weakness of this system is the potential for the results of the oracle to be manipulated through collusion. If any entity controls more than 50% of the total votes, they can essentially set the median unilaterally.
Therefore, we have mitigation methods for the main attack vectors we anticipate — Sybil attacks and collusion:
- We use Gitcoin Passport to prevent one person from creating multiple accounts.
- Crowd-workers can stake before starting a task. This system is inspired by the peer prediction method and the reward increases according to the correlation of the “Skin in the Game” amount. Conversely, the staking could be slashed if the correlation is too low.
- We plan to calculate the social distance between Crowd-workers by mixing the addresses participating in DeCartography as Crowd-workers into the analysis target of DeCartography.
Academic research has shown that such bridging-based mechanisms can help incorporate higher-quality responses and reduce the risk of a single community colluding to manipulate the results of the oracle.
Crowd-workers who have a track record of sending high-quality responses have a high “common sense” score, while new Crowd-workers who do not yet have a track record (have not completed many tasks) have a low “common sense” score. The reward paid by DeCartography for the work of end-users is determined according to this “common sense”. Accounts with high common sense scores are more likely to receive high rewards. For more information, please see the reward determination mechanism.
It is still unclear how a completely new idea like DeCartography will work in which version, and many iterations of experiments will likely be needed to clarify what rules yield good results in various contexts. We appreciate your feedback.
