digital trade fair | fan yuan: build data circulation and utilization infrastructure to solve the problems of "inability to supply, inability to circulate, poor use, and weak security"
2024-09-26
한어Русский языкEnglishFrançaisIndonesianSanskrit日本語DeutschPortuguêsΕλληνικάespañolItalianoSuomalainenLatina
on september 25, at the data factor governance and marketization exchange event of the 3rd global digital trade fair, fan yuan, chairman of anheng information, delivered a keynote speech entitled "reflections and practices on data security, compliance, circulation and utilization", elaborating on anheng information's in-depth thinking and rich practices in data circulation and utilization infrastructure and ai-enabled data security protection for the leaders and guests on site.
fan yuan said that with the booming digital economy, data elements have become basic elements, and their value and influence are increasing day by day. data must be protected and utilized, which undoubtedly puts higher requirements on data security and privacy protection. the current data element market as a whole is still facing the realistic problems of "inability to supply, unsmooth flow, poor use, and weak security". data infrastructure is born out of necessity and has epoch-making significance. how can we empower the activation of data value in the whole society by building data infrastructure?
fan yuan used three practical cases to illustrate ahnheng information’s thinking and practice on data circulation and utilization infrastructure.
as a major financial district in shenzhen, futian district has a good foundation for technological innovation, but it still faces the problem of poor data circulation. based on the premise of "original data does not leave the domain, data is available but not visible", ahnheng information has effectively solved the problem of poor data circulation between various government commissions and bureaus, superiors and subordinates, data supply and demand between government and enterprises, and trust, security and privacy protection between data supply and demand parties, and has promoted the efficient internal and external circulation of government public data. at the same time, it also combines blockchain technology to ensure that highly sensitive public data can be managed, controlled and traced.
previously, in the medical artificial intelligence competition hosted by the national health commission and organized by the zhejiang provincial health commission, ahnheng information developed a large model training and reasoning platform based on confidential computing technology on the basis of the privacy computing platform, becoming the first platform in the country to pass the evaluation of the enhanced trusted execution environment for large models by the china academy of information and communications technology. in the large model training and reasoning scenarios, the platform has five core capabilities, including security protection, security assessment, model development, low-cost migration and acceleration device support, which can be applied to the two scenarios of large model fine-tuning and secure reasoning.
fan yuan introduced that anheng information continues to explore the field of data element circulation supervision, develops "data compliance circulation digital certificates", and deeply participates in the construction of "three numbers and one chain" in china digital valley. the core of the three numbers and one chain data infrastructure is to create a cross-domain interconnected data resource discovery network, and then coordinate the data delivery platform based on privacy computing and blockchain to achieve cross-industry, cross-regional and cross-domain data trusted circulation. in august 2024, the practical case of the "three numbers and one chain" data infrastructure in china digital valley was selected as a recommended case for the national data bureau's digital economy innovation development pilot zone.
at the same time, fan yuan combined his practical experience over the past year and said that the big model technology has had a revolutionary impact on data security. ai allows data security to better understand data, understand scenarios, and be more accurate, achieving full-process security control.
in the field of ai-enabled data security, data classification and grading are the foundation of data security work. ahnheng information has created a data classification and grading model of hengnao (big model) + aisort, which has increased the efficiency of data classification and grading by 30 times. ahnheng information has also become the first manufacturer in the industry to implement classification and grading products in combination with big models. the combination of ahnheng information's api risk monitoring system and hengnao's vertical security big model will increase the efficiency of alarm analysis by 200% and the efficiency of security operations by at least 250%, achieving a disruptive security operation effect of "from manual rule setting, manual analysis to automatic analysis".
in fan yuan's view, the successful practice of ai-enabled data security is inseparable from the technical support of the hengnao security vertical domain big model. as a typical case of making full use of data and exploring the value of data, anheng information hengnao 2.0 has been upgraded to an intelligent form. empowered by hengnao, based on a new generation of ai-driven intelligent and scalable security architecture, it supports the implementation of scalable work and collaboration modes, and can be extended to multiple security scenarios. among them, data security api security, data classification and grading, dlp and other scenarios are all industry (and even global) firsts.
