In the era of the data economy, data is becoming more important to businesses.Whether it is a technology company that needs algorithm model training, or a traditional industry that depends on data for research and development (such as pharmaceutical companies), it needs data to support business development.However, since data is prone to the hidden dangers of C-side privacy invasion and B-side data leakage during the use of data, it is increasingly urgent to use technical means to solve the need for efficient flow and use of data under the premise of legal compliance.36 A company recently contacted by 锘 崴 Technology, through the creation of 锘 崴 trust privacy cloud computing platform, to achieve the effect of data sharing and analysis in a secure environment.Specifically, this platform uses software and hardware encryption computing technology (such as multi-party secure computing, homomorphic encryption, trusted computing environment), alliance computing technology, blockchain traceability technology, and customizable hyper-converged infrastructure technology.Protect the rights and interests of data owners, data consumers and data regulators under the legal compliance framework.Solve the problem of privacy protection in the entire process of storage, transmission and calculation, and ensure the integrity and authenticity of the calculation process.The company is currently focusing on medical, financial and other fields, of which the medical field is the focus of the current stage.The two founders of Dr. Wang Shuang and Dr. Zheng Zheng both had rich scientific research and work experience in the fields of privacy computing and biomedical information in the United States.It is also during these experiences that Wang Shuang and Zheng Zheng found a pain point in the use of data in the medical field, that is, data cannot be circulated smoothly-data scattered in various hospitals could not be effectively transferred due to concerns about privacy leaks and loss of data ownership,Sharing and analysis has hindered the development of precision medicine to some extent.For example, the current demand for secure use of big data in some hospitals lies in data compliance and data collaboration.Doctors need sufficient medical data (including medical record samples, etc.) as support when carrying out scientific research work, but due to concerns about risks such as privacy leaks or non-compliance, doctors may not be able to get enough data.And the platform can make the data “available and invisible”-that is, the original data is encrypted and shared and calculated, and only relevant results such as customer reports are fed back.concern.Specifically at the landing level, the types and formats of medical big data vary.Medical data can be roughly divided into three types: genomics data, image data and clinical data.Genomics and imaging data is a relatively standard format, and the company’s system can support the access and analysis of related formats.The clinical data needs to be adjusted and processed by the customer according to the relevant international standard model, and then connected to the system.In the actual situation, it is not realistic to push the data one by one. It is not realistic for customers to standardize the data. Therefore, I choose to cooperate with the national team of the medical group. These organizations have their own licenses and have adopted relevant data format standards.The data into the system.When the medical group selects a technology provider, it mainly considers the technical level of the bidding company from the dimensions of calculation methods and data carrying capacity.In addition, to really land and use, in addition to the availability and reliability of the technology, the product also needs to be easy to use.For example, in a business application scenario, customers are not necessarily doctors or scientific researchers who have been specially trained. The method is to provide customers with different backgrounds with a simple operating system. Customers only need to insert data into the system.Operate through the visual interface, select the visualization scheme you need and finally form a data report.Beyond the software level, 锘 崴 is currently developing and testing related hardware systems, which will be combined with existing systems in the form of coprocessors to provide more efficient privacy protection computing effects.It will be provided to customers as part of the next-generation total solution after the iteration is completed.In the field of big data privacy, two traditional secure computing methods are more common, including providing a safe house sandbox and providing data desensitization.Because the former adopts the isolation method that still requires administrators or users to participate in it, the latter still cannot completely avoid the leakage of sensitive information by eliminating sensitive fields. Therefore, neither of these methods can guarantee the absolute security of data.The Wei Wei system can also be empowered to other fields without the underlying technology being changed. For example, customers in the financial industry will have data model collaboration requirements. Some private equity funds can provide trading strategies to the party for encryption. After encryption,The model can be used to empower third parties (such as customers), and the strategy owner does not have to worry about the leakage of trading strategies, and can also generate profits through other cooperation methods.In terms of business model, the current plan is to cooperate with data resource-type partners and provide technical services. The other party provides the corresponding data and customer resources. Both parties will jointly tap the market and make profit distribution together.In the future, other customers may also explore business models such as commissioning and distribution.To achieve the purpose of “making data available and invisible,” technically requires the dual blessing of algorithms and engineering capabilities.The team believes that the team ’s founders have previously fully understood the advantages and disadvantages of different algorithms in this field through the work of the global privacy and secure computing competition established in the United States, and can combine different types of algorithms into high-performance hybrid secure computing solutions.Program.And its scheme has been deployed in some hospitals on the west coast of the United States to undergo stress tests, which provides a prerequisite for subsequent commercial landing.锘 崴 Tech believes that the company’s future industry barrier to privacy computing lies in the data network.At present, the company first obtains customers and data sources through algorithmic barriers and industry experience, and gradually forms its own data indexing platform in the application. “For example, the index of our platform will eventually help pharmaceutical companies accurately locate a province and city under the premise of privacy protection.Whether a hospital in China can provide the data needed for new drug development or whether a big data company can provide the credit information required by a bank, and the platform can further seamlessly support subsequent privacy computing needs after providing retrieval services.”Wang Shuang introduced.In addition, because of the need to provide data indexes, I will score and provide data pricing in the system based on the quality and availability of the data.If the quality of the data is better, the corresponding availability will be higher, and there will be more opportunities to participate in more important research.At the same time, some data providers can also continuously optimize data according to the purpose of data use, forming a virtuous circle to improve the data quality of the entire industry.At present, there are companies in the industry that conduct business on privacy computing, including Huakong Qingjiao and Digital Technology.Among them, Huakong Qingjiao’s team is from Tsinghua University. It mainly researches, develops and operates big data security fusion technologies, standards and platforms based on modern cryptography and game theory.The company’s current main product is the PrivPy standard platform, which can meet the needs of a wide range of user groups to protect multi-party data privacy and realize collaborative computing.Data Science and Technology, through data science and engineering, cryptography (multi-party secure computing, differential privacy, etc.), federal learning and other technologies, helps companies collaborate on secure and private data collaboration. Its founder, Song Yimin, once worked at Facebook Ads..
Create a privacy cloud computing platform
36 氪 starterAutopilotBlack technologycapital marketcareducationEntrepreneurshipForefrontFrontlinegameGo publicInformationInternet of ThingsinvestmentNet red economyNew businessNew economyOccupationreal estatestockTechnologyTelematicsthe Internetunmanned