[Speaker Info]
==========
Heyi Wu graduated from Southeast University majoring in network security, has rich security work experience, has participated in the construction of multiple central ministries and commissions-level network security systems, as well as offensive and defensive drills and actual combat. End-to-end implementation of research and products in AI security, data security, cloud security, security development, intelligence hunting and other fields. Delivered speeches at well-known network security conferences at home and abroad, such as BH MEA, HITB, BSides, Kanxue SDC, XCon, Geekpwn, etc., participated in the compilation of multiple security standards at home and abroad (IEEE P2841, China self-developed operating system security, etc.), published many EI/SCI papers and patents , has a number of security certificates at home and abroad (CISP, CCSK, CDPSE, etc.), and has submitted several 0day vulnerabilities, general security issues of self-developed linux systems and advanced attack hunting traceability reports to the regulatory authorities.
[Abstract]
==========
Chinese government and enterprise organizations will replace domestic operating systems and application software on a large scale. However, most security vendors are currently unable to cover this area. Moreover, the domestic terminal security software has weak capability in detecting and killing malicious ELF samples. Some hackers and attack teams have targeted products that use domestic operating systems, and the existing detection capabilities are easily bypassed by attackers. Worse still, there is no established channel for exchanging malicious ELF samples among vendors. The situation is: There are very few domestically produced ELF malicious sample databases; There is zero work on the generation of adversarial malicious software targeting ELF; Currently, the existing work on countermeasures against PE malicious samples cannot generate executable binary files; there is no work on AI-based generation of malicious ELF samples. Thus, a system that generates adversarial malicious samples in the domestic environment using AI technology is needed to establish an advantage in the malicious sample library, and enhance the capability to detect and kill malicious software. We develop such a system and solve the above problems.
