USENIX Courses

Confusing Value with Enumeration - Studying the Use of CVEs in Academia

Explore how academic security research misuses CVE identifiers, revealing that 34% of claimed vulnerabilities lack confirmation and examining misconceptions in the review process.

• YouTube
• 19 minutes
• Self-Paced
• Free Video

POPS - From History to Mitigation of DNS Cache Poisoning Attacks

Discover POPS, a novel DNS cache poisoning prevention system that integrates with IPS to detect and mitigate statistical attacks with zero false positives using simple rules and TC flag enhancement.

• YouTube
• 17 minutes
• Self-Paced
• Free Video

DNS FLaRE - A Flush-Reload Attack on DNS Forwarders

Discover DNS FLaRE, a cache-based timing attack exploiting DNS forwarders to infer website visits and IoT device activity through resolution time discrepancies.

• YouTube
• 16 minutes
• Self-Paced
• Free Video

EmbedX - Embedding-Based Cross-Trigger Backdoor Attack Against Large Language Models

Discover EmbedX, a novel cross-trigger backdoor attack using embedding vectors to manipulate large language models while maintaining stealth and preserving utility.

• YouTube
• 16 minutes
• Self-Paced
• Free Video

FABLE - Batched Evaluation on Confidential Lookup Tables in 2PC

Discover FABLE, a system for efficient batched evaluation on confidential lookup tables in secure two-party computation, achieving up to 393× speedup over existing methods.

• YouTube
• 15 minutes
• Self-Paced
• Free Video

SoK: Inaccessible and Insecure - An Exposition of Authentication Challenges Faced by Blind and Visually Impaired Users in State-of-the-Art Academic Proposals

Explore authentication security and accessibility challenges for blind and visually impaired users through systematic evaluation of 50 academic schemes using the ALESA framework.

• YouTube
• 15 minutes
• Self-Paced
• Conference Talk

I Know What You Said - Unveiling Hardware Cache Side-Channels in Local Large Language Model Inference

Uncover hardware cache side-channel vulnerabilities in local LLM inference that expose token values and positions, compromising user privacy through novel eavesdropping attacks.

• YouTube
• 15 minutes
• Self-Paced
• Free Video

Dumbo-MPC - Efficient Fully Asynchronous MPC with Optimal Resilience

Discover Dumbo-MPC, a breakthrough in asynchronous multi-party computation offering optimal resilience and guaranteed output delivery with efficient dual-mode offline protocols.

• YouTube
• 10 minutes
• Self-Paced
• Free Video

Stack Overflow Meets Replication - Security Research Amid Evolving Code Snippets

Explore how Stack Overflow's evolving code snippets impact security research stability and learn recommendations for treating it as time series data.

• YouTube
• 18 minutes
• Self-Paced
• Free Video

SoK: Understanding zk-SNARKs - The Gap Between Research and Practice

Explore the comprehensive analysis of zk-SNARKs, bridging theory and practice by examining 40+ schemes, 11 libraries, and identifying key gaps between research and industry implementation.

• YouTube
• 18 minutes
• Self-Paced
• Free Video

TORCHLIGHT - Shedding Light on Real-World Attacks on Cloudless IoT Devices

Discover how TORCHLIGHT exposes anonymous Tor-based attacks on cloudless IoT devices, revealing 45 vulnerabilities including 29 zero-day exploits affecting 12.71 million devices worldwide.

• YouTube
• 13 minutes
• Self-Paced
• Free Video

AirTag-Facilitated Stalking Protection - Evaluating Unwanted Tracking Notifications and Tracker Locating Features

Explore AirTag anti-stalking features through research evaluating unwanted tracking notifications and locating capabilities across iOS and Android platforms.

• YouTube
• 11 minutes
• Self-Paced
• Free Video

From Constraints to Cracks - Constraint Semantic Inconsistencies as Vulnerability Beacons for Embedded Systems

Discover NÜWA, a static analysis technique that detects embedded system vulnerabilities by identifying constraint semantic inconsistencies, uncovering 152 new security flaws.

• YouTube
• 13 minutes
• Self-Paced
• Free Video

My ZIP Isn't Your ZIP - Identifying and Exploiting Semantic Gaps Between ZIP Parsers

Discover how ZIP parser inconsistencies create security vulnerabilities across 50 implementations, enabling attacks on email gateways, documents, and applications.

• YouTube
• 12 minutes
• Self-Paced
• Free Video

GPUHammer - Rowhammer Attacks on GPU Memories are Practical

Discover the first successful Rowhammer attack on NVIDIA GPUs, demonstrating how to exploit GDDR6 memory vulnerabilities to inject bit-flips and compromise ML model accuracy by up to 80%.

• YouTube
• 14 minutes
• Self-Paced
• Free Video