On the evening of August 12th to 15th, Beijing time, the DEF CON 30 CHINA Party was simultaneously unveiled at the Xirang Yuanverse World and the offline venue in Beijing. It also presented the real-time broadcast of DEF CON 30 in Las Vegas, USA. A series of core links such as KEYNOTE, VILLAGE and CONTEST are shown to domestic geeks.
is the top professional conference in the field of global network security. Since Baidu Security and DEF CON jointly introduced it to China in May 2018 and founded DEF CON CHINA, DEF CON CHINA has actively promoted the construction of China's network security community and empowered the growth of Chinese geeks. Baidu Security's exploration and achievements in the security field, especially the AI security field, have been demonstrated in previous activities.
In-depth exploration of three major dimensions and layout of ecological development in the field of AI security
Adhering to the concept of "AI is safer", Baidu Security has always been committed to cutting-edge exploration in the field of AI security. In 2020, Baidu Security took the lead in proposing AI security research in three dimensions: Security (strong adversarial security), Saftety (non-adversarial security), and Privacy (data and privacy security), which is groundbreaking in the field of industry AI security.
is deeply involved in cloud technology, and the AI security solution matrix escorts the AI era.
provides AI model attack detection and adversarial defense, model robustness reinforcement and systematic evaluation, complex scene security verification, AI deep forgery detection, and full life cycle data. To explore security, privacy-protecting machine learning and other technologies, Baidu Security has launched multi-industry solutions such as cloud security, AIoT security, data security and privacy protection, intelligent connected car security, and anti-black industry to comprehensively safeguard the AI era. Major business scenarios of cloud security. In the field of cloud security, based on core technologies such as AI, big data, cloud computing and ,
provides enterprises with a commercial security product and service system integrating intelligence, network, host, application, auditing and security services; and for enterprises Different industries and enterprises have formed personalized and scenario-based solutions to help enterprises build cloud security systems and service ecosystems.
In the field of AIoT, Baidu Security provides technology-driven "cloud-pipe-end + compliance access" full-stack security solutions, covering mainstream and cutting-edge AIoT industries such as smart home, smart driving, smart finance, and smart medical care. At present, it has cooperated with more than 100 domestic leading and innovative companies such as Skyworth , Konka, Inspur , Apollo, and Xiaodu.
In the field of data security and privacy protection, Baidu Dianshi privacy computing platform relies on a series of privacy computing technology accumulation such as secure multi-party computing, federated learning, confidential computing, and secure digital sandbox to create data security covering the entire life cycle of data. and privacy protection solutions to promote digital, networked and intelligent transformation of enterprises.
In the field of intelligent connected vehicle security, Baidu Security quickly locates security concerns at each stage for manufacturers from all dimensions and throughout the life cycle, helping companies reduce costs, increase efficiency, and manage compliance risks. It currently covers the three major business lines of smart transportation, autonomous driving, and smart cockpit, dozens of businesses, and has successively empowered more than ten car companies and T1 suppliers. In the field of anti-black industry,
is based on threat intelligence big data and Baidu Brain AI knowledge map. Baidu Security carries out risk labelling, interception prompts and search blocking for harmful information on the entire network, ensuring that harmful information reaches users from the source. In 2021, Baidu safely blocked more than 45.45 billion malicious web pages, and intercepted fraudulent websites and apps that reached users 31 million times.
"AI Security + Metaverse" join forces to enter the field of autonomous driving safety
Based on continuous innovation and practice in the three dimensions of AI safety research, Baidu Security's latest launch is the independently developed autonomous driving safety research project "Tiema Glacier". This is a security test platform focusing on immersive high-fidelity scenarios and rich in AI confrontation. With the help of a large number of high-fidelity test scenarios, it has passed the three evaluations of "vehicle-in-the-loop", "hardware-in-the-loop" and "software-in-the-loop" The solution can realize safety evaluation of indicators such as autonomous driving visual robustness and Al safety confrontation in various customized test environments .
At the same time, we integrate the metaverse into these evaluations to combine physics and virtuality in a more realistic way. In the process of interaction between highly simulated scenes and test vehicles, we create a random situation that is closer to the real road test effect and reduce the testing complexity. Comprehensive cost. The fuzzing fuzz testing concept that "Iron Horse Glacier" relies on has also been a "frequent guest" at many top global network security conferences including DEF CON in recent years.
Currently, "Tiema Glacier" has been put into practice in application scenarios such as autonomous driving model safety evaluation and intelligent cockpit interaction design. The project has also been successfully launched on Baidu's first domestic Yuanverse product " Xirang ".
It is worth mentioning that Baidu Security BCTF and University of California, Irvine entered the DEF CON 30 Contest site to jointly organize BCTF·AutoDriving. In this DEFCON, the difficulty of the competition questions has been upgraded. Based on the five scenarios of attack, evidence collection, detection, vehicle collision, and intelligent planning, a new safety-critical scenario Traffic Sign is added. We look forward to the wonderful performance of this year's contestants.
With an open attitude, we jointly build a new line of defense for AI security
As the "Oscar" in the security field, DEF CON CHINA has always been unswerving in promoting the construction of China's network security community. "DEF CON is an event that only focuses on technology and community, rather than commercialization. It is also an effective promotion for people in the non-security industry to understand and learn." Ma Jie, vice president of Baidu, said that as network security becomes more and more important, More and more people are paying attention to it, and the geek community is becoming more vibrant. Maintain enthusiasm, maintain love, and explore the most cutting-edge and most imaginative things. This is also the responsibility of geeks.
On the one hand, by participating in security conferences such as DEF CON, the public can consciously understand many new network security issues and challenges in the intelligent era; on the other hand, practitioners in the security field can also obtain available tools to solve existing or Hidden danger problems, thus, safety problems of all sizes can be effectively controlled and solved.
Ma Jie called on: “Security is not a matter for any one individual, one company or one community. We must unite and collaborate with an open attitude and common concepts to make the community more prosperous and jointly maintain the security ecology in the AI era. "
pioneers openness and remains true to its original aspirations. With the help of DEF CON CHINA, a large-scale event with industry benchmark significance, Baidu Security looks forward to continuing to maintain in-depth cooperation with partners from industry, academia and research, so that AI security can be rooted in more fields and make this new line of AI security defense more unbreakable!
