- ARPA participates in preparing The Institute of Electrical and Electronics Engineers Standards Association (IEEE) P2830 standard.
- ARPA joins the group led by Alibaba and Alipay, with representation from Baidu, Lenovo Group, and Zhejiang University.
ARPA, the privacy-preserving computation network, has announced that it’s participating in preparing The Institute of Electrical and Electronics Engineers Standards Association (IEEE) P2830 standard. The standard covers how to train machine learning algorithms using data from multiple sources and has just entered the ballot stage of the IEEE Standard Association Standard Development Process. The working group in which ARPA is participating is led by Alibaba and Alipay, with representation from Baidu, Lenovo Group, and Zhejiang University, among others.
Over recent months, blockchain privacy has become a hot topic, as developments such as multiparty computation, such as ARPA uses, and zero-knowledge proofs have gained prominence. The shift has come about in no small part due to the booming cryptocurrency and DeFi markets. As the sums involved have increased exponentially, there’s increasing concern that large-value transactions can be tied back to wealthy individuals, creating security and privacy risks.
Furthermore, investment is pouring into the blockchain privacy segment. PayPal recently acquired Curv, a custody provider using multiparty computation (MPC) technology, and ZenGo recently raised $20 million in funding to further develop its non-custodial wallet.
Multiparty computation has proven to be a critical development in privacy-preserving technologies. It’s based on the principles of Shamir’s Secret Sharing, where a piece of private data is split and shared between many participants, who can’t discern the original data from only their piece. It’s becoming more frequently deployed in institutional-grade custody solutions. However, ARPA incorporates MPC for use across its entire blockchain platform.
A Universal Standard for MPC in Machine Learning
Now, IEEE P2830 standard aims to introduce standardization into the process of using MPC in machine learning. The IEEE (Institute of Electrical and Electronics Engineers) Standards Association focuses on development of standards for the IEEE, which are used by engineers and developers all over the world. The submission of P2830 has been performed by the working group led by Alibaba and with participation from ARPA and other representatives from industry and academia.
According to the announcement from ARPA, the standard “refers to the required practices for training a model using encrypted data aggregated from multiple sources and processed by a third-party trusted execution environment.”
The IEEE Standards Development Process specifies six stages. The P2380 standard has successfully completed the first three and has now entered the ballot stage. Entering this stage is confirmation that the standard has reached a stable stage following the initial drafting. The standard must successfully achieve a 75% turnout from the balloting group to pass the ballot. Further, 75% of ballots must provide approval. Once this stage is passed, the standard is submitted to the Institute for final approval.
ARPA has made considerable progress since the inception of the project in 2018, and is well-positioned to support the establishment of standards governing the development of multiparty computation. For example, ARPA has worked with a group of multi-national clients in the finance sector to develop a blacklist program that allows them to assess risks across their entire combined client base, but without sharing sensitive data about the identity of each firm’s clients. It has also helped in shaping the first set of national standards for Multi-party computation in China.
The project provides a cross-industry platform for analyzing data from multiple sources in a way that’s private. It covers any scenario where regulations, competition, or ethical considerations prevent data from being openly analyzed in aggregate.
- New York Bill Proposes 3 Year Crypto Mining Ban
- EGLD Holders Can Now Compute and Claim MEX Tokens
- May The Crypto Be With You: QAN to Launch Star Wars NFTs
- Miami Bitcoin Mining: the Best Crypto Mining Capital in the Future
- Latvian Art Center Becomes The First Of Its Kind To Purchase KIWIE’s NFT Digital Art Piece