On Jul.2, 2020, the Head of PlatON Community Kai Yu, as the special guest of Industrial Blockchain Talks hosted by Milin Finance, talked about the solutions that PlatON’s privacy-preserving computation provides, the use cases, and how the powerful combination of privacy-preserving computation and blockchain facilitates the data-element-oriented market.
What Solution Can Privacy-Preserving Computation Provide for Privacy Protection in Data Sharing and Data Exchange?
Yu: Generally speaking, privacy-preserving computation is a technology that executes the computing collaboration without disclosing the data and computing methods to all parties get involved in the computing.
Proposed in 2016, privacy-preserving computation includes cryptography, Trusted Execution Environment (TEE), Federated Learning, etc., while cryptography includes Homomorphic Encryption (HE), Multi-Party Computation (MPC), Zero-Knowledge Proof (ZKP), etc.
If there are two traditional financial institutes that want to know the whitelist and blacklist of credit from each other without unpacking their own credit data, privacy-preserving computation can be applied for solving this dilemma.
What’s the Biggest Challenge to Realize Multi-Party Data Sharing with Data Privacy Kept?
Yu: The biggest challenge is that privacy-preserving computation can’t be applied in all scenes due to the existing technology bottleneck, which means it may take time to realize the completely common use of privacy-preserving computation. Performance would be the key to cause that challenge. Plus, the traditional concept about data security also has an effect on the spread of privacy-preserving computation. People usually think the data ownership and use of right can’t be separated.
Data is valuable. Thus, we are dedicating to promoting privacy-preserving computation and having it popularized in all industries that are sensitive about data privacy. By doing so, privacy-preserving computation can be the real infrastructure for the whole digital era in the near future.
Does PlatON Have Any Good Solution to Balance Security, Efficiency and Data Silos of Privacy Protection?
Yu: As the privacy-preserving computation of PlatON is based on cryptography, I will share my views based on cryptography.
Currently, cryptography technologies such as HE, ZKP, MPC, etc. have been recognized by the market, however, there are still some bottlenecks and weakness. For example, the communication of MPC requires high bandwidth, and ZKP needs massive computing power to improve its computing speed. What PlatON keeps investing are focusing on MPC, then improving the algorithms, lowering the communication volume, and transferring the consumption from communication to computation. This is actually PlatON’s solution to balance the security, efficiency and data silos when dealing with privacy protection. As for ZKP, we have a hardware team to improve the computing performance using hardware, and that’s the reason why hardware is a part of the Grants program we launched in early 2020.
As the requirements vary and become more and more complicated, we will keep improving the algorithms to make our privacy-preserving computation solutions fit more and more use cases.
How Do You Think About the Relationship between Privacy-Preserving Computation, AI and Blockchain?
Yu: Privacy-preserving computation based on cryptography fits AI well. As there are more and more strict rules to control data disclosure, it’s harder and harder for AI projects to get data resources, bringing in challenges for the development of AI industry. Privacy-preserving computation is just the solution. Thus, we launched Rosetta, a privacy AI framework that combines AI algorithms and privacy-preserving computation technologies to help developers easily get their hands dirty on that.
As for blockchain and privacy-preserving computation, actually, they are independent while can empower each other. Privacy-preserving computation protects the data privacy for blockchain, while blockchain, as the underlying support infrastructure for distributed economies, provides payment and settlement functions.
Can You Share the Use Cases Where PlatON’s Privacy-Preserving Computation are Applied in the Financial Industry?
Yu: The first one is the union query of credit. We provided the blockchain-based MPC solution to offer the customizable computing logic template and multi-party access. Our solution can output the required query results without data disclosure, encrypt the original data and store it on blockchain system. It can meet all kinds of audit requirements.
The another one is Supply Chain Finance. We built a new financial model of supply chain finance with information symmetry sharing and transitive core enterprise credit value for upstream and downstream firms. And we provided the solution of which the digital assets can be confirmed, proceed and transferred based on blockchain and cryptography algorithms, and realized asset registration, right confirmation of asset, asset digitalization, digital asset management, regulation & audit, and multi-layer privacy protection.
Besides Financial Industries, Are There Any Other Industries Need to Be Improved by Privacy-Preserving Computation and Blockchain?
Yu: I think industries that are sensitive about data privacy all need to be improved by privacy-preserving computation combined with blockchain, for example, the union data analysis of research institutes which can be applied in medical data analysis, physical examination for military affairs, and quality trace of outsourcing in the aerospace industry to ensure all data synced on the chain including the design, component materials, production link, experimental data and acceptance report are tamper-resistant and traceable.
As data privacy demands get more and more urgent, PlatON are working on the use cases in finance, join credit, joint marketing, intelligent manufacturing, biomedical, etc. with privacy-preserving computation combined with blockchain.
What Are the Most Crucial Problems for the Development of Privacy-Preserving Computation?
Yu: Though privacy-preserving computation has been recognized by more and more industries, it’s still in the early stage. According to the existing challenges I mentioned before, I think the most crucial problems for the development of privacy-preserving computation are:
-
We all must be clear that the development of technology can’t be independent from business practice. With efficient business practice, we can find more and more real use cases;
-
We are looking forward to the related laws like GDPR to limit the data disclosure taken place in firms and institutes, by which we can step into the complete digital era with the backup from privacy-preserving computation;
-
We need more and more talents specialized in cryptography and privacy-preserving computation to push the development of the industry.