Tags → #artificial intelligence

ByzSFL Enhances Secure Federated Learning with Zero-Knowledge Proofs

Recent research introduces the DuoAgg framework for secure federated learning, which enhances data privacy and resilience in collaborative environments through innovative methodologies and advanced encryption techniques.
SecAlign Develops Defense Against Prompt Injection Attacks

Recent research by Sizhe Chen et al. introduces **SecAlign**, a multi-faceted approach to enhance the security of large language models against prompt injection attacks while maintaining utility and user privacy.
AI Innovations Enhance Cybersecurity Software Quality and Security

Recent research highlights innovative methodologies that utilize Artificial Intelligence to enhance vulnerability detection and prioritization in cybersecurity.
Autonomous Vehicles Vulnerable to Electromagnetic Signal Attacks

Recent research highlights the vulnerabilities of autonomous vehicles to Electromagnetic Signal Injection Attacks and emphasizes the need for improved threat models and multi-sensor integration to enhance security.
PhishAgent Introduces Advanced System for Phishing Detection

PhishAgent is a new multimodal phishing detection system designed to enhance cybersecurity defenses through advanced detection mechanisms and real-time adaptability.
Strategies to Enhance Security of Large Language Models

Recent research highlights innovative strategies, such as Layer-specific Editing and Reinforcement Learning from Human Feedback, aimed at enhancing the security of large language models against adversarial threats while maintaining their performance.
Backdoor Attacks Identified in Android Deep Learning Models

Recent research highlights vulnerabilities in deep learning systems, particularly regarding backdoor attacks in mobile applications, emphasizing the need for enhanced model protection and user education.
LENS-XAI Framework Advances Intrusion Detection in Cybersecurity

A recent study on the LENS-XAI framework for anomaly detection highlights its potential to improve intrusion detection systems while addressing challenges such as resource constraints and the need for enhanced explainability in AI systems.
Study Examines MrDeepFakes Sexual Deepfake Marketplace Dynamics

A recent study examines the MrDeepFakes community, highlighting the risks of deepfake technology in sexual exploitation and the need for regulatory frameworks and detection tools to address its misuse.
AI and Cybersecurity: Privacy, Consent, and Innovations

The intersection of artificial intelligence and cybersecurity is evolving, highlighting the need for enhanced privacy measures and user-centric consent management systems to ensure data security and compliance.