The Cryptonomist conducted an interview with Eric Vreeland, CSO at Polyhedra Network to talk about Web3, zero knowledge (ZK) proof and artificial intelligence (AI).
Can you provide an overview of Polyhedra Network’s mission and vision, particularly in the context of Zero-Knowledge (ZK) proofs, and how you see it transforming the web3 and web2 ecosystems?
Our vision is to unlock exponential improvements to computational power and cross-platform interoperability through the use of zero-knowledge (ZK). While zero-knowledge proofs (ZKP) are most commonly known for their privacy characteristics, the ability of ZKPs to compress large datasets into small proofs has huge potential to revolutionize computation. Instead of needing to verify every individual computation or transaction, a ZKP can batch these together so that they can be quickly checked by a verifier. This could lead to massive performance improvements in computationally intensive applications like AI and machine learning.
As traditional financial institutions continue to explore blockchain technologies, the ability to provide verifiable computation and to preserve data privacy will make zero-knowledge an area of growing interest.
Until now, the biggest hurdle for zero-knowledge adoption has been the complexity of the underlying technology along with the limitations in terms of performance (speed). Polyhedra has continued to invest heavily in ZK research and continues to push the limits as to how quickly zero-knowledge proofs can be generated. Our most recent proof system, Expander, is currently the world’s fastest proof system and makes zero-knowledge applications much more practical. Our focus now is on making development with these highly performant proof systems more accessible so that any developer can utilize the benefits of zero-knowledge.
How do Zero-Knowledge proofs enable secure, trustless transactions, and what are the primary benefits of these transactions for users and developers in both web3 and web2 environments?
Zero-knowledge facilitates secure and trustless transactions by enabling one party to prove the validity of a statement to another without revealing any underlying information. For developers, this reduces the need for trust-mechanisms or third-party intermediaries which leads to a reduction in costs.
In web3, ZKPs allow for the offloading of computations and minimize on-chain data. Computation on-chain is expensive and handling the majority of compute off-chain results in significant cost reduction. In web2, zero-knowledge improves data security and privacy allowing for trust-minimized interactions without the need for intermediaries.
In what ways do ZK proofs contribute to system scalability and operational cost reduction, and can you share specific examples or use cases where these benefits are most apparent?
Zero-knowledge proofs contribute to system scalability and operational cost reduction by enabling efficient verification of large computations and data sets without requiring full access to the underlying information.
This reduces the computational load on primary systems and minimizes data transfer, leading to faster processing times and lower infrastructure costs. Specific examples include zk-rollups in blockchain technology, where numerous transactions are processed off-chain and only a concise proof is verified on-chain, significantly enhancing throughput and reducing gas fees.
In web2 environments, ZKPs are used in privacy-preserving data analysis, allowing companies to perform computations on encrypted data without compromising user privacy, thereby cutting down on the need for extensive data handling and security measures.
What are the key advantages of trustless cross-chain bridges, and how does Polyhedra Network leverage these bridges to enhance interoperability and security across different blockchain networks?
The key advantage is the elimination of additional trust-assumptions or third parties. Alternative bridges most commonly use either a validator network or middle-chain to confirm the state of one chain and relay that to another.
With zero-knowledge this can be eliminated by simply generating a proof of the state of one chain and verifying that proof on the other chain. Removal of these additional trust-assumptions decreases attack surface area and minimizes trust, requiring only trust in the data and consensus of the blockchain itself.
How can the integration of ZK proofs with blockchain technology improve the integrity and reliability of AI models, and what potential applications do you foresee in this intersection of technologies?
Zero-knowledge improves the integrity and reliability of AI models by providing tamper-proof verification that a model’s training processes were performed correctly without exposing the underlying data. This gives end-users confidence that models were built on trustworthy data, improving the user-experience and enhancing credibility in what is commonly a very opaque field.
The tamper-proof aspect of zero-knowledge also significantly reduces the opportunity for fraud. Applications with sensitive data like financial services or healthcare are industries where zero-knowledge could provide an immediate benefit.
Can you discuss some of the recent developments at Polyhedra Network and share any upcoming projects or innovations that you are particularly excited about? What future directions do you envision for the network in the ZK space?
We are the authors of the zkBridge whitepaper and creators of www.zkbridge.com the most widely used zero-knowledge secured bridge. We connect over 25 blockchains and have facilitated over 20MM cross-chain transactions.
Recently we are working on Proof Cloud, a cloud proving platform that decreases costs and increases the efficiency of the proof generation process. This will be open to the public at the end of the month.
Lastly, we have open-sourced our record-breaking proof system Expander. Orders of magnitude faster than alternative proof systems, developers utilizing Expander will see incredible improvements in both the speed and cost of the proof generation.
Over the next 12 months we plan to dedicate substantial resources to the exploration of how zero-knowledge can be applied to AI and machine learning with a number of AI specific developer tools and applications in the pipeline.