360 security large model released, Zhou Hongyi: The real opportunity for large models in the future lies in the enterprise market

On August 9, 360 (601360.SH, hereinafter referred to as "360") Group released the first deliverable security industry model in China - "360 Security Model".

According to reports, the 360 ​​security large model is based on 360's self-developed cognitive general-purpose large model "360 Smart Brain", combined with 360's past 15 years of AI security applications and security big data to train a vertical large model of the security industry. During the privatization deployment process of 360 security large model, it will match the "enterprise security intelligent control system", and use the intelligent central scheduling model, knowledge base, special plug-in, and combine with the private domain data of the enterprise, so as to realize security question answering experts and security operation experts There are two major forms to assist in the safe operation of enterprises. At present, these have been implemented in 360's internal and self-owned products.


According to the data, 360 is an early enterprise in China that deploys artificial intelligence large-scale models. 360 Intelligent Brain Large-scale Model already has ten core capabilities in China, covering all application scenarios of large-scale models. Zhou Hongyi judged that the real opportunity for big models in the future lies in the enterprise market. At present, relying on the "360 Smart Brain" capability, 360 has released an enterprise-level AI large-scale model solution, following the four principles of "safety, benevolence, credibility, and controllability" to create an enterprise-level vertical large-scale model, which is the first in the taxation and corporate service industries Landed, and both were selected as "Top Ten Typical Scenario Cases of General Artificial Intelligence Large-scale Model Industry Application in Beijing". It is understood that Qihoo 360 launched a security industry model this time, hoping to cloudify and intelligentize national security capabilities.

Zhou Hongyi introduced at ISC 2023 that 360 mainly develops large-scale models of the security industry along three directions. The first is to be a "safety know-it-all". Through a large amount of security knowledge and data training, the general-purpose large-scale model is trained as a "security expert"; It is an auxiliary offensive and defensive decision-making. When the system encounters an attack alarm, the large model will judge whether it is an attack or a false alarm; the third is to combine the large model with 360’s existing network-wide security brain to improve the security service effect. The release of the 360 ​​security model is an important achievement of the 360 ​​exploration model to empower the security industry.

Some people in the industry pointed out that many companies that make large-scale models do not understand security, and companies that understand security do not have the ability to make large-scale models. 360 is just a cross-border between the two. 360 has the world's largest network security big data. Relying on these data training It is logical to come up with a large model of the security industry.

It is reported that based on the brand-new concept of "Security as a Service", 360 also simultaneously launched a multi-tenant cloud-based security service platform - 360 Security Cloud, fully opening up 360's national-level security capabilities and building security infrastructure and public service facilities on the cloud. , and then realize eight security services.

 

Guess you like

Origin blog.csdn.net/GZZN2019/article/details/132216906