OpenAI's SORA has captured the public's attention, demonstrating the transformative power of AI and its ability to reshape our world once again. In the field of cryptocurrency and blockchain, decentralized AI projects such as Gensyn, OORT, and Bittensor continue to emerge, accelerating AI development by leveraging the advantages of decentralization (enhanced data privacy and cost savings). Using blockchain technology and crypto-economic incentives, decentralized AI encourages global participants to contribute computing power and data, promoting innovation and widespread adoption of AI technology.
This article will introduce one of the most fundamental protocols for decentralized AI: Proof of Honesty (PoH). Specifically, it explores how to incentivize geographically distributed service providers (aka nodes) to contribute to the global optimal goal and verify decentralized resources (such as bandwidth, computing power, and storage space) to ensure that they operate as promised, aiming to build truly trustworthy AI.
PoH incentivizes nodes to take actions to optimize social goals
For efficiency and fairness, networks should prioritize users with higher utility needs, such as what we do in Internet bandwidth allocation and cellular scheduling systems.
However, the lack of comprehensive global information makes it challenging to optimize resource allocation in a decentralized environment. From a game theory perspective, nodes allocate resources based on local knowledge and self-interest, leading to suboptimal social outcomes and limited network performance. The PoH protocol leverages blockchain technology to incentivize service providers with cryptocurrencies proportional to their network contributions.
For example, in decentralized AI, the datasets for AI models can be stored in multiple geographically distributed service providers. In other words, service providers typically store fragments of various datasets from different users in the network. Suppose a service provider receives a file download request from one user in the network.
When a user obtains all fragments of a dataset from these service providers, the cryptocurrency rewards will be shared proportionally by the service providers who contributed the dataset fragments. By doing so, the PoH consensus protocol incentivizes service providers to cache frequently accessed dataset fragments nearby and deploy them in areas with high and reliable bandwidth. Therefore, all service providers are independently working towards the global optimal goal; that is, PoH optimizes network topology and resource allocation in a decentralized manner.
PoH realizes decentralized trusted computing
Security and trust are critical in the digital age, especially for AI computing tasks outsourced to unknown service providers in decentralized AI networks. PoH is a vigilant guardian of the AI computing landscape, ensuring that outsourced tasks are completed accurately and honestly. It employs a mechanism inspired by the use of entrapment techniques by law enforcement, carefully incorporating test tasks or "phishing tasks" into actual tasks. If providers try to cut corners, "PoH officers" (some service providers are designated as officers who create and distribute phishing tasks in the network) catch them in the act of using these phishing tasks. This deters dishonest behavior and ensures the process is secure without the need for constant supervision.
In particular, PoH embeds "phishing tasks" within regular tasks, making them indistinguishable from real tasks, but with the expected results known to the outsourcer. This strategy creates uncertainty for service providers about which tasks are tests, encouraging honesty to avoid detection and penalties. By not disclosing phishing tasks, providers are motivated to consistently do their best because they know that any task could be a test and that the consequences of getting caught outweigh the benefits of cheating. PoH's mathematical model calculates the optimal number of phishing tasks to include, balancing security needs with computational efficiency. The model takes into account computational costs, the likelihood of dishonest behavior, and the impact of erroneous computations on the integrity of the entire system. The goal is to maximize network security while minimizing unnecessary overhead, guaranteeing both the security and efficiency of the system.
Democratizing AI development
Although still in its infancy, decentralized AI has the potential to revolutionize the way AI is developed. By facilitating direct interactions between developers and users, bypassing the traditional gatekeepers of centralized institutions, it paves the way for a more democratic and trustworthy AI ecosystem.
This shift could significantly alter the power dynamics of the AI market. As more vendors adopt decentralized AI, the dominance of proprietary models could wane, leading to a significant reduction in market control. As a result, this paradigm shift is expected to usher in an era of increased transparency and inclusivity in AI development, marking a substantial step toward democratizing access to AI technologies.
About Oort
OORT is a decentralized data cloud platform that aims to integrate global computing and storage resources, maximize privacy and cost-effectiveness, and provide an enterprise-grade, decentralized solution for generative artificial intelligence and data-driven businesses.
OORT's core technology is backed by U.S. patents and academic publications, and has established solid partnerships or customer relationships with more than 40 projects/companies and more than 20 prestigious U.S. universities (e.g., Harvard, MIT, University of Chicago, Columbia University, Duke University, University of Michigan, Northwestern University, New York University, etc.).
40+ channels, ecosystems, and customers:
BNB Chain-BNB Greenfield (Binance), DEII (Dell), Lenovo Image (Lenovo Image), Tencent (Tencent), Alibaba Cloud (Alibaba), Seagate (Seagate), Akamai (Akamai), Storj, Poolz Finance, OORT SWAP, iPolloverse, Grown, Crust, Sinso, Parallell, Revolution NFT, Coresky NFT, Mises, Appsco, NumenCyber, SPEARHEAD, Layerr, Uverse, Y0 NFT, ThreeFold, IOV Labs, blue world, Crypto Legacy, Supermoon Camp, Walken, COTX, Clocr, Midas, iCrypto World, Verida, 404DAO, Midasreless, Supermoon Camp, 1024bank, etc, FilSwan, AFRICA2.0 (Africa 2.0), Ivy League (Ivy League), Together Labs (its IMVU has nearly 7 million monthly active users), Finstreet (Indian crypto education platform - with 4.5 million community members and nearly 100,000 students), etc.
Investment agency:
EMURGO Ventures (ADA), Aves Lair, Microsoft, Google, Redline DAO, Sanctor Capital, Waterdrip Capital, LinkVC, etc.#热门话题
