Building decentralized knowledge using incentives.
Synapse is a HackRice project by Noor El Sonbaty, Matt Joss, Yorke Rhodes IV, and Achal Srinivasan.
Students around the world use collaborative learning environments to succeed at their universities. Though, a host of extenuating circumstances make it more difficult for certain students to build a deep understanding of course materials. Consider, for instance: student-athletes required to miss class frequently, non-native students who have difficulty understanding a professor’s presentation style or notation, or international students with highly disparate educational backgrounds.
Peer learning can help close this educational gap, but typically this exchange of class notes and similar resources occurs only offline or amongst a close-knit group of students. As a result, a student’s success in the classroom is in part dictated by their intimacy with fellow classmates. Naturally, a shared, online resource which bubbles up a simple, accurate understanding of course materials from a pool of relevant students would greatly benefit the community. Unfortunately, this collaborative effort is difficult to organize, susceptible to free-riding, and provides no clear incentive for contributions.
Existing shared-knowledge platforms fail to incentivize their contributors and are vulnerable to centralized attack vectors.
This case study is indicative of a larger problem with collaboratively-built knowledge-sharing platforms like Wikipedia. Due to a lack of meaningful incentives for subject matter experts, few of them actively contribute to public resources. Rarely will experts spend invaluable time creating a widely useful explanation of their knowledge, and seldom continue to participate in the public discourse in an effort to maintain quality after their initial contributions. Furthermore, platforms like Wikipedia are subject to manipulation and cooption due to their centralized nature. As a result, documents on shared-knowledge platforms cannot realistically be used for academic citations or as reliable evidence of a contributor’s domain expertise.
Reallocating control of content editing from a central organization to contributors immunizes shared-knowledge platforms against content poisoning.
If a platform like Wikipedia could somehow only allow past contributors to make decisions on new changes, the content could not be changed from its original intent by an external actor. This create an issue for would-be contributors, though. How can one submit content for the first time, and what would motivate this decision?
On the other hand, the organizations which support shared-knowledge platforms are unlikely to relinquish control over content editing. They have too much reputational risk to allow unvetted contributors democratic control over content that is directly associated with their organization and brand. Only a decentralized platform could avoid this reputational risk, but the implementation of this type of platform would introduce various security risks.
Enter Ethereum: a decentralized platform that runs programs called smart contracts exactly as intended with minimal possibility of downtime, censorship, fraud or third-party interference. This is secured by a blockchain, a shared global infrastructure that represents transferal and ownership of digital value. Ethereum is a substrate-shifting technology stack that enables decentralized applications with designs never before plausible. For a more detailed explanation of Ethereum, visit the project website.
Synapse, a HackRice project, is a decentralized knowledge-sharing platform with token-based contribution incentives, built on top of Ethereum.
At the highest level, Synapse is composed of ecosystems, which define some domain of knowledge. Anyone with an Ethereum account can create an ecosystem, like ‘Systolic Algorithms’ or ‘Siberian Cats’. This ecosystem defines an area of expertise that the creator identifies as corrupt, biased, or nonexistent among other ecosystems. Each ecosystem has attached content (shared knowledge) which is publicly accessible, and a unique token central to the incentivization scheme.
An ecosystem’s token represents a relationship between individual Ethereum accounts and their respective token balances. A more apt description of this token might be a badge or claim, because it can be accumulated but not necessarily exchanged. Accordingly, this token has 2 primary functionalities:
- minting to a specified Ethereum account
- retrieving the balance of a specified Ethereum account
Within an ecosystem, the token plays a simple but powerful role.
Each time a new change is proposed to an ecosystem’s content, a time-sensitive decision must be made concerning the merge of these changes into the canonical version of the content. To reach consensus on approval or rejection of a proposal before it expires, a vote takes place among ecosystem contributors. A contributor’s influence on the outcome of the vote is proportional to their token balance. To elaborate…
If Alice casts an approval on proposal P targeting ecosystem X, proposal P’s approval tally is increased by Alice’s balance of ecosystem X’s token. When proposal P expires, if the approval tally is greater than the rejection tally, ecosystem X’s content is updated with proposal P’s changes.
This raises the question: how does one acquire an ecosystem’s token to participate in voting and content curation?
Synapse defines only one way for an ecosystem’s token to be accumulated: earning through good contributions. When an ecosystem is created, the creator is minted the entire supply (1 unit). When a proposal is approved, this token is minted again (1 unit) to the proposal author, rewarding a good contribution. This completes the system which limits content curation to historical contributors.
Outside of an ecosystem, the token plays a more nuanced role.
Observers of an ecosystem are exposed to two readily available resources: canonical content and token balances. Recall that the token balance of a specific Ethereum account represents the quantity of approved contributions within the ecosystem. Now consider that some ecosystems might be more easily contributed to than others, which can be described with a difficulty metric. This metric helps to distinguish ecosystems which have different approaches to the same (or similar) domains, competing for mind-share of readers and active contributors. There are two concrete factors which inform an observer’s perceived difficulty of an ecosystem:
- quality of content — proposed contributions must be of equal or greater quality than the canonical quality to get approved
- criteria for voting — proposed contributions must satisfy a majority of voters’ criteria for approval
As a result, overcoming an ecosystem’s difficulty repeatedly takes significant skill. Theoretically, this skill would be directly correlated with a contributor’s subject matter expertise. This skill is measured by the contributors’s token balance, because it is incremented for each approved contribution. In other words, an observer’s ecosystem-specific evaluation of content quality and contribution criteria dictates the significance they attach to an account’s ecosystem-specific token balance, which represents the skill of a subject matter expert.
The combined effects of an ecosystem token’s internal function and external significance incentivize contributors.
The democratization of the collaboration process via an ecosystem vote solves many of the exploitable problems inherent in centralized knowledge-sharing platforms. It does not, however, explain how Synapse introduces incentives for contributors beyond the public good.
In our modern information society, there is clear demand for subject matter experts. Third party observers such as potential employers looking to hire, academic researchers making citations, or even venture capital investors (think ICOs) need a skill metric for the supposed expert they interface with. In general, this is the idea behind LinkedIn skill endorsements and similar online badges. Typically, this metric informs a decision about how much time, effort, or value it is worth to spend in an interaction with the individual. The problem with these types of skill attestations is that they can easily be spoofed within a group of colluding individuals or with a sybil attack.
On Synapse, an individual’s skill value is tied to how much good work they have completed in a given ecosystem. Parties who desire SMEs find them by parsing published portfolios. If a Synapse contributor publishes their skill value, a subjective analysis of the associated ecosystem’s difficulty determines the contributor’s externally perceived worth as a subject matter expert. Rational SMEs seeking to capitalize on their competitive knowledge advantage will attempt to monetize their contract with third parties, to the extent that their perceived worth permits. This eliminates “spoofing” by shifting third-party trust from the skill value itself to the skill-backing ecosystem’s content and members. Therefore, contributors are incentivized to complete good work on the Synapse platform because they can extract rent. By publishing their ecosystem-specific token balances, contributors expose themselves to further financial opportunities.
Future Work
If an ecosystem’s content is to be useful, it must always have some canonical state which represents historical collaborative efforts. Within the EVM, this canonical content exists only in hashed form to avoid storage costs of variable length content. Therefore, an ecosystem’s associated content updates if and only if the canonical content hash changes on-chain.
Consequently, there is no way to extract content directly from the Ethereum network; a storage layer which can be authenticated using the on-chain hash is necessary for readers.
A centralized organization (like Wikipedia, that aims to be an encyclopedia) could feasibly create a list of trusted Synapse Ecosystems whose content they could host on-site. These types of websites would serve as valuable indexing tools for end users, defining some navigational hierarchy or graph for moving between the content of different ecosystems (see Wikipedia Portals). These hosting services would also play an invaluable role for the Synapse platform, ensuring storage redundancy of the content which derives an ecosystem’s current (or past) content-hash stored on-chain.
Readers who wish to evaluate the integrity of a centrally hosted page’s content, with respect to the Synapse Ecosystem’s network consensus, could use an interface for comparing the on-chain content hash to a computed hash of the page’s content. As a result, the reader is empowered to vet content providers based on adherence to the consensus of a trusted Synapse Ecosystem’s network.
There would also be impetus for a decentralized autonomous organization to assume the role of these aforementioned centralized organizations. This DAO could democratically decide on their list of trusted Synapse Ecosystems using a Token Curated Registry. Stakeholders (i.e. contributors) within Synapse ecosystems would have an incentive for their ecosystem to be included on this registry to improve their externally-facing Synapse portfolios. The ability for this DAO to establish a brand and perceived network effect would create demand for the token intrinsic to this TCR. The resultant deflation of this token could theoretically fuel decentralized hosting of content (on IPFS, for instance).
Conclusion
A decentralized platform like Synapse can provide an opportunity to level the playing field for students who experience educational inequities, by incentivizing genuine and meaningful participation to collaborative educational materials. The Synapse hackathon team are now all members of our on campus blockchain organization: Rice Blockchain. We are excited about novel applications of blockchains like Synapse which demonstrate potential to help resolve societal inequities.
View our source code here. Start a discussion below, and please reach out to us at blockchain@mailman.rice.edu with any questions or concerns!
Special thanks to Achal Srinivasan for contributing to this post.
