Cybersecurity
Overview
Cybersecurity studies how to prevent, detect, investigate, and recover from attacks on software, networks, organizations, and cyber-physical systems. Common approaches include signature-based defenses, statistical and deep-learning intrusion detection, graph and temporal analysis, formal verification, threat-intelligence mapping, and reinforcement-learning or language-model agents. Work spans settings such as cloud and IoT networks, industrial control systems, connected vehicles, vulnerability assessment, and security operations, with benchmarks increasingly testing realistic attack chains and operational response. A central tension is that strong results on fixed datasets or simulated environments may not transfer to changing, adversarial deployments, where false alarms, safety, cost, and human oversight also matter.
Latest papers
379 papers21 papers in the last four weeks, down 46% on the four weeks before. 0.2% of all new papers.
CyberArena: Benchmarking AI Agents in Live Symmetric Attack-and-Defense Combat
Autonomous Cyber Operations Gym for Training and Evaluating AI Models for Defence
AutoSec-RAN: Specification-Grounded LLM Fine-Tuning for Automated Security Analysis of O-RAN
Game-Theoretic Reinforcement Learning based Topology-Agnostic Defense for Virtual Power Plants
SEEK: Agentic Knowledge Graph Retrieval for Secure Code Generation
Defensive Sufficiency in a Stackelberg Model of AI Security
Temporal transformer CAN encoder with federated lightweight heads for anomaly detection
Contextualization of Third-Party Cloud Security Findings
Evaluating Behavioral Context for Interpretable IAM Policy Risk Scoring in Cloud Environments
From Sandbox to Enforcement: Confidence-Qualified Threat Intelligence for Critical Infrastructure
Polar: LLM-Powered Synthesis of Real-World Cyber Evidence for Prioritization and Mitigation
Does AI Help Cyber Attackers or Defenders? Evidence from Nonpublic Vulnerabilities and Subsequent Attacks
Adaptive-Shot Hybrid Quantum Anomaly Detection for Tactile Internet Security: Reliability-Aware Measurement Allocation Under Resource Constraints
CyTReX: Explainable AI-Based Cybersecurity Threat Reasoning Framework for DER Networks
A Quantitative Analysis of Graph Representation Strategies for Cyber Attack Detection
CVE2AP: Automated Generation of PDDL-Encoded Attack Paths via Large Language Models
GPT-5.5 offers the best quality-cost trade-off and error feedback yields the most consistent quality improvement.