Will Privacy computing give birth to the next trillion giant?

Amazon, Twitter, Google, Alibaba are definitely the first names that rushed into our minds when it comes to giants in the Internet era. They deal with unimaginable large amounts of data every day. Our behavior is collected day by day in their cloud computing, which makes their data models more intelligent and complete, covering almost all differences kinds of crowds.

The “different kinds” mentioned here are extremely subdivided concepts. For example, “the new London father, in the middle of the economic level, who wants to take the CFA test”. The pre-attribute of each concept is very long. The AI library can accurately push advertisements, shopping recommendations, blogger recommend to you based on the choices of other people similar.

At this point, from a personal point of view, it is difficult to define whether this matter is good or bad. Although everyone knows to protect their privacy, the “protection of privacy” will not affect our happy surfing, nor has it become a pain point for individual needs. We even more addicted to it based on analyze and recommend algorithms. Therefore, we provide more data to the other party with the recommended content unknowingly.

From the perspective of enterprises. Protecting consumer data privacy is not only to protect their own business secrets, but also a responsibility they need to fulfill. However, only relying on the data collected by the enterprises themselves can no longer meet their needs for higher-precision data models. Business owners are eager to share data, but they are subject to the problem of data leakage during sharing.

Privacy computing was born.

What is privacy computing?

The “Privacy Computing Research Scope and Development Trend” released in 2016 defines it as follows: Privacy computing is the computing theory and method for the full life cycle protection of private information.

It is the Computable model and axiomatic system for privacy measurement, privacy leakage cost, privacy protection, and privacy analysis of complexity when the ownership, management and use rights of private information are separated.

Talk like a human being, from a business perspective, privacy computing can help companies perform functional operations on data or jointly train a model without knowing the plaintext of the data, so as to help themselves make better decisions. From a personal perspective, privacy computing can help us protect the privacy of any behavior so that these private data can bring us benefits.

However, privacy computing is still subject to technological progress currently, only perform relatively simple calculations.

Okay, I will use the most easy-to-understand language to introduce to you several technical directions in the field of privacy computing now.

Several mainstream technologies for privacy computing

The concept of Federated Learning is proposed by the Google team in 2016. After multiple data participants reach a consensus on training the same model, they regularly upload the model gradients and train the model while the data will not be transferred out of the local devices, so as to protect data privacy while using each other’s data.

Secure Multi-Party Computation is a set of technologies based on cryptography, and the famous “Millionaire Problem” put forward by Mr. Yao Qizhi is often appeared with it. Multiple participants input information to jointly calculate a function. For example, 10 millionaires get a financial ranking without knowing each other’s property.

Trusted Execution Environment is a hardware-level technology that allows each data participant to perform calculations in the same trusted hardware device. The performance of this method is very efficient and can almost reach the performance of plaintext calculation, but there is also a certain risk that you have to completely trust the hardware equipment vendor.

Privacy computing is unstoppable

Privacy computing has been mentioned many times in 2020, and many people predict that 2021 will be the first year of the explosion of privacy computing & blockchain.

Blockchain has been trying to solve the problem of trust since its inception, and the separation of trust covers data verification, data circulation, data security, and privacy protection.

Blockchain is expected to become an indispensable option in privacy computing products. Data security, compliance, and reasonable and effective use can be realized on the basis of ensuring data credibility.

  1. The blockchain can guarantee the end-to-end privacy of private computing task data.
  2. The blockchain can guarantee the security of the entire life cycle of data in private computing.
  3. The blockchain can guarantee the traceability of the privacy calculation process.

The combination of blockchain and privacy computing allows the original data to achieve collaborative computing and data privacy protection among multiple nodes, without the need for collection and sharing. At the same time, it can solve the problems of excessive data collection, data privacy protection, and single point of data storage leakage in the big data model.

The blockchain ensures the credibility of the calculation process and data, and the privacy calculation realizes that the data is available but not visible. The two combine with each other and complement each other to achieve a wider range of data collaboration.

Undoubtedly, privacy computing will be the infrastructure that will operate in the era of big data in the future, allowing data ownership to truly return to individuals and serve more companies.

The popularity of private computing may represent the freedom of society in the future and become a new “social freedom indicator”.