Security Comparison Hub
13 papers - avg viability 4.8
Current research in security is increasingly focused on addressing vulnerabilities in complex systems, particularly as the landscape evolves with new technologies. Recent work on searchable symmetric encryption highlights how system-level monitoring can expose previously unconsidered leakage patterns, emphasizing the need for robust defenses against sophisticated attacks. In the realm of privacy, studies on Tor website fingerprinting reveal that real-world conditions significantly affect the effectiveness of these attacks, prompting a reevaluation of existing security measures. Additionally, the exploration of backdoor attacks in federated learning and graph neural networks underscores the necessity for layer-aware detection strategies, given the unique vulnerabilities these decentralized models present. Meanwhile, advancements in post-quantum cryptography are being assessed for their practical implementation challenges, particularly on different hardware architectures. Collectively, these efforts reflect a shift towards a more nuanced understanding of security threats, aiming to bridge theoretical gaps and enhance the resilience of systems against emerging attack vectors.
Top Papers
- Improved Leakage Abuse Attacks in Searchable Symmetric Encryption with eBPF Monitoring(7.0)
Leverage eBPF monitoring to expose system-level leakages in Searchable Symmetric Encryption, enabling stronger leakage abuse attacks and highlighting the gap between theoretical security and practical system exposure.
- Reality Check for Tor Website Fingerprinting in the Open World(7.0)
A tool to detect website fingerprinting attacks on Tor networks, enabling enhanced security measures for privacy-focused applications.
- SoK: Evolution, Security, and Fundamental Properties of Transactional Systems(7.0)
A security analysis framework for transactional systems, identifying vulnerabilities and proposing an extension to the ACID model.
- PQC-LEO: An Evaluation Framework for Post-Quantum Cryptographic Algorithms(7.0)
PQC-LEO is a benchmarking suite that automates the evaluation of post-quantum cryptographic algorithms, enabling developers to optimize performance across different architectures.
- From Data Leak to Secret Misses: The Impact of Data Leakage on Secret Detection Models(5.0)
A tool to identify and mitigate data leakage in AI-based security models, ensuring accurate performance evaluation.
- Next-Gen CAPTCHAs: Leveraging the Cognitive Gap for Scalable and Diverse GUI-Agent Defense(5.0)
Develop a scalable framework for Next-Gen CAPTCHAs to defend web systems against advanced GUI-enabled agents.
- Poisoning the Inner Prediction Logic of Graph Neural Networks for Clean-Label Backdoor Attacks(5.0)
Developed a method to enhance clean-label graph backdoor attacks by poisoning the inner prediction logic of Graph Neural Networks.
- Exploiting Layer-Specific Vulnerabilities to Backdoor Attack in Federated Learning(4.0)
Develop robust defenses in Federated Learning to mitigate layer-specific backdoor attacks like LSA.
- Image-based Prompt Injection: Hijacking Multimodal LLMs through Visually Embedded Adversarial Instructions(4.0)
Develop an Image-based Prompt Injection tool to expose and address vulnerabilities in multimodal language models.
- A Systematic Literature Review on LLM Defenses Against Prompt Injection and Jailbreaking: Expanding NIST Taxonomy(3.0)
Defend against prompt injection and jailbreaking in LLMs with a comprehensive mitigation strategy catalog.