As of today, decentralised applications face challenges in securely organising and automating data in off-chain contexts. Enter HyperOracle, presenting zkOracle — a solution that leverages zero-knowledge proofs to bridge on-chain data with off-chain environments seamlessly.

HyperOracle acts as the channel for bringing machine learning on-chain, unlocking scalable AI capabilities while maintaining decentralisation and verifiability.

At the crossroads of AI and blockchain, transparency is crucial. Despite assurances of openness, the opacity of AI models remains a significant concern. Ethereum’s founder, Vitalik Buterin, sees potential in the convergence of AI and crypto, where blockchain’s transparency offsets AI’s inherent obscurity.

Endorsed by the Ethereum Foundation, recognising its advanced work with Halo2 technology HyperOracle has made significant strides in blockchain standards and technology, contributing to Ethereum (ERC-6150, ERC-7007), Celestia, and Uniswap.

In today’s article, we delve into HyperOracle’s role in democratising AI, bridging the gap between transparency and innovation, and shaping the future of decentralised applications.

In today’s AI and ML sectors, transparency is a pressing concern. Despite claims of openness, understanding the specific models behind AI outputs remains a challenge. Vitalik Buterin sees promise in combining AI with crypto, as crypto’s transparency can counteract AI’s opacity. HyperOracle steps in with a promising solution.

HyperOracle aims to develop the World Supercomputer, a global network securely linking diverse peer-to-peer networks. This initiative enhances Ethereum’s capabilities, supporting computational demands and scaling storage. By integrating AI and ML on-chain, HyperOracle plans to revolutionise smart contracts with intelligence, leveraging blockchain’s transparency.

While Ethereum enables on-chain computing, its limitations hinder fully on-chain ML/AI. HyperOracle addresses this with opML and zkML, innovative technologies validating computations on-chain, overcoming challenges in cost and performance. zkOracle, a decentralised oracle network, utilises zero-knowledge proofs to bring on-chain data off-chain, perform off-chain computation on-chain, and verify off-chain proofs, ushering in a new era of transparency and efficiency.

Ethereum serves as the consensus ledger, while the zkOracle network runs resource-intensive computations, connecting various components with zero-knowledge proof technology.

opML optimises on-chain AI and ML, offering efficiency, scalability, and decentralisation. By embracing this novel approach, they pretend to unlock the full potential of AI and blockchain, maintaining transparency, security, and ethical responsibility.

As established in their documentation, HyperOracle is a programmable zkOracle protocol that powers smart contracts with arbitrary computing and richer data sources. HyperOracle offers full security and decentralisation for trustless automation and on-chain AI/ML so builders can easily create next-gen dApps.

In the HyperOracle network, multiple zkOracle nodes can be described as a node, similar to an Ethereum node, that carries out computations but also produces zero-knowledge proofs.

As you can imagine, the technical aspects of this project are quite complex. Even after delving into the documentation in detail, understanding it completely can be challenging. Nevertheless, let’s dissect the revolutionary elements that HyperOracle introduces, which allows us to better grasp its future potential.

zkMLinference is HyperOracle’s trustless machine learning inference protocol based on zero-knowledge proofs. It is a zkAutomation-like input and output zkOracle Meta App that integrates zk machine learning frameworks using zkGraph as the core component to enable on-chain ML inference. Developers will be able to use zkMLinference to build any DApp driven by ML and secured by Ethereum.

zkML empowers blockchain technology to discreetly authenticate AI model applications, ensuring data confidentiality throughout the entire process.

zkML is an exciting intersection of Crypto and AI, with Modulus Labs and Giza Tech leading the way in its commercial application. Despite its promise, there’s a noticeable gap in the AI supply chain, particularly in authentication, which hasn’t been fully addressed. The slow adoption of zkML frameworks is mainly due to their lengthy proving times and high computational costs, highlighting the need for further development in this area.

To bridge the gap between zero-knowledge proofs and machine learning models, two distinct implementation approaches have been suggested:

As mentioned above, zkML technology is not yet fully optimised. HyperOracle acknowledges this and, while they are working on it, their founder introduced a new concept called Optimistic Machine Learning (opML) as a solution until the launch of zkML.

As the creators of the first open-source implementation of opML, they believe that opML can facilitate the integration of AI and Crypto by eliminating cryptographic overhead through game theory.

opML conducts AI model inference and training/fine-tuning on-chain using an optimistic verification mechanism. Compared to zkML, opML provides ML proofs with low cost and high efficiency. Additionally, opML has low hardware requirements, allowing it to run large language models on common PCs without GPUs.

opML employs a verification game (similar to optimistic rollup systems) to ensure decentralisation and verifiability of the ML compute process.

HyperOracle stands as the first open-source implementation of opML.

zkGraphs are like customisable “Smart Contracts” for HyperOracle nodes, managing data behaviours and zero-knowledge proof generation, similar to how smart contracts define Ethereum node computations. Smart contract developers can construct both smart contracts and zkGraphs, allowing users to interact with both.

A zkGraph comprises three main components:

For deployment, all ZKgraph code files are stored in EthStorage, a decentralised storage scaling layer supported by Ethereum ESP.

To understand its direct competitors better, let’s examine what this project brings to the table, as it innovatively applies well-known technologies. Firstly, HyperOracle introduces a blockchain service, the zkOracle protocol, for performing complex calculations in smart contracts, akin to Chainlink’s approach. It also employs zkGraph technology for data processing post-collection, similar to The Graph. Also, it’s important to note that HyperOracle isn’t the only project offering AI features on the blockchain, even within the zkML area, other projects, such as Modulus Labs, are notable players.

HyperOracle stands out from traditional knowledge graphs such as The Graph by utilising zkWASM. Some of its benefits include:

Compared to traditional knowledge graphs, which often lack decentralised security due to their reliance on conventional indexing methods, HyperOracle offers distinct advantages. Traditional approaches may depend on DAOs or legal documents, adding uncertainty to automation processes. Additionally, their security is typically less robust, often relying on multi-signature dispute councils for conflict resolution.

To simplify, what they suggest is that Chainlink and The Graph are networks that fetch and integrate data into blockchain apps but can’t fully ensure data accuracy. Chainlink uses multiple data sources and security measures which, although substantial, are not foolproof. Both rely on data aggregation and incentives to aim for correctness, yet this doesn’t guarantee absolute data integrity. HyperOracle believes more is needed for secure, trustless applications, and that’s why they are developing a network that uses mathematical proofs to verify data quickly, offering a more reliable solution than traditional trust-based or incentive-driven methods.

Thoughts: Keep in mind that the innovative features often mentioned in the whitepapers of new projects can be challenging to realise in practice. In contrast, traditional oracle solutions such as Chainlink have a track record of proven effectiveness over time. The ability of these new features to match or exceed such established standards remains to be seen.

This area becomes more complex due to the novelty of this emerging trend and the protocols developing it. For instance, Modulus Labs is combining AI with blockchain using ZK proofs and machine learning to produce transparent, verifiable AI algorithms for blockchain applications. It adopts a unique strategy, especially in handling the demands of proof generation, while promising to allow Dapps to incorporate AI and ML without compromising trust or security.

Another significant project in this domain is Gensyn, a blockchain-based AI compute protocol that recently completed a $43 million Series A funding round led by a16z.

At Greythorn, we make it a priority to stay updated on the latest developments in blockchain and crypto technology. One particularly intriguing area of growth lies in the intersection of AI and blockchain. While traditional smart contracts follow preset actions triggered by events, the incorporation of AI and ML enables smart contracts to autonomously make real-time decisions. These proactive capabilities represent exciting additions to blockchain functionality.

If you’re keen on exploring this topic further, we invite you to check out our latest research on OLAS.

For more insights and upcoming research pieces, follow us on LinkedIn, X, and Medium. You can also visit our website for additional information.

If you find this information valuable, we welcome any interaction with this post.

Looking forward to connecting,

The Greythorn Team.

This presentation has been prepared by Greythorn Asset Management Pty Ltd (ABN 96 621 995 659) (Greythorn). The information in this presentation should be regarded as general information only rather than investment advice and financial advice. It is not an advertisement nor is it a solicitation or an offer to buy or sell any financial instruments or to participate in any particular trading strategy. In preparing this document Greythorn did not take into account the investment objectives, financial circumstance or particular needs of any recipient who receives or reads it. Before making any investment decisions, recipients of this presentation should consider their own personal circumstances and seek professional advice from their accountant, lawyer or other professional adviser. This presentation contains statements, opinions, projections, forecasts and other material (forward looking statements), based on various assumptions. Greythorn is not obliged to update the information. Those assumptions may or may not prove to be correct. None of Greythorn, its officers, employees, agents, advisers or any other person named in this presentation makes any representation as to the accuracy or likelihood of fulfilment of any forward looking statements or any of the assumptions upon which they are based. Greythorn and its officers, employees, agents and advisers give no warranty, representation or guarantee as to the accuracy, completeness or reliability of the information contained in this presentation. None of Greythorn and its officers, employees, agents and advisers accept, to the extent permitted by law, responsibility for any loss, claim, damages, costs or expenses arising out of, or in connection with, the information contained in this presentation. This presentation is the property of Greythorn. By receiving this presentation, the recipient agrees to keep its content confidential and agrees not to copy, supply, disseminate or disclose any information in relation to its content without written consent.