The 3rd AEGIS Symposium on Cyber Security

The Annual rEading Group in Information Security (AEGIS) Symposium is an annual symposium held online by a group of researchers of Privacy and Security in Computer Science.

In 2023, the 2nd AEGIS Symposium hosted 12 talks given by a variety of security researchers in security, covering a wide range of topics with a highlight on LLM security. The 2024 symposium will be the 3rd symposium of the AEGIS symposia series.

As usual, the 3rd AEGIS Symposium will be delivered virtually in Zoom. For any additional questions, please contact the organization committee.

Important News:

We have finalized the presenters this year, featuring 12 talks in four sessions. The topics of the presentation will be finalized in mid June. Looking forward to see you in AEGIS 2024!

Upcoming Event:

Abstract:

Rust's rise as a systems programming language has reshaped expectations for memory safety and performance. Yet, its unique features, such as iterators, traits, and compiler optimizations, introduce novel challenges for reverse engineering, especially for security-critical applications like malware analysis and firmware hardening. This talk serves as a call to action for the research community, highlighting the hurdles posed by Rust binaries and the gaps in existing analysis tools. It introduces the ongoing exploration of probabilistic binary analysis and large language models (LLMs) as potential solutions. By sharing early-stage findings, this presentation aims to inspire innovative approaches to tackle the complexities of Rust binary analysis and beyond.

Sessions:

• Session 1: Machine Learning and Large Language Model
• Session 2: Security Threat Analysis and Modelling
• Session 3: Network Security
• Session 4: Invited Keynotes from Database and Software Engineering

Presenters:

Shufan Zhang	A Private Talk: New Models, Results, and Systems for Privacy-Preserving Database Query Processing	University of Waterloo
Xinyue Shen	Emerging Attacks in the Era of Generative AI	CISPA Helmholtz Center for Information Security
Zhuo Zhang	Beyond C/C++: Motivations and Challenges in Rust Binary Reverse Engineering	Purdue University
Ling Zhang	Efficient Query Processing in Resource Constrainted Settings	University of Wisconsin-Madison
Yichen Li	Your Code Secret Belongs to Me: Neural Code Completion Tools Can Memorize Hard-Coded Credentials	Chinese University of Hong Kong
Shixuan Zhao	Reusable Enclaves for Confidential Serverless Computing	The Ohio State University
Limin Wang	Model-Driven Attack Detection Methods	Nanjing University
Yan Long	How Sensors Leak Your Secrets: Protecting Security and Privacy of Sensing in Modern Computer Systems	University of Michigan
Qifan Zhang	ResolverFuzz: Automated Discovery of DNS Resolver Vulnerabilities with Query-Response Fuzzing	University of California, Irvine
Yunyi Zhang	Rethinking the Security Threats of Stale DNS Glue Records	Tsinghua University
Weiheng Bai	APILOT: Navigating Large Language Models to Generate Secure Code by Sidestepping Outdated API Pitfalls	University of Minnesota

Organization Committee

Program Chair	Yuqing Yang	The Ohio State University
Program Co-Chair	Xiang Li	Nankai University