Anomaly Detection with LLMs
Research on using LLMs for zero-shot or few-shot detection of anomalies, outliers, and unusual patterns across various domains
Anomaly Detection with LLMs
Research on Large Language Models in Anomaly Detection with LLMs
Zero-Shot Anomaly Detection with MLLMs
Detecting anomalies without prior training data
Detecting Advanced Cyber Threats with LLMs
Using AI language models to identify stealthy persistent threats
Visualizing Time Series Anomalies
Leveraging Multimodal LLMs for Enhanced Anomaly Detection
Sherlock: AI-Powered Video Anomaly Intelligence
Advanced LLM system for detecting and localizing abnormal events across multiple scenes
Zero-Shot Video Anomaly Detection
Customizable surveillance without retraining or specialized expertise
Efficient Log Anomaly Detection with LLMs
Using Parameter-Efficient Fine-Tuning for cost-effective security monitoring
Smart Flood Detection Without Labels
Using Zero-Shot Learning & Bayesian Modeling for Urban Flooding
Smarter Log Anomaly Detection with LLMs
Leveraging LLaMA2 to enhance cybersecurity through intelligent log analysis
Improving AI Security Through Multi-Modal Detection
Enhancing out-of-distribution detection with cross-modal alignment
Detecting Ghost Behaviors in LLMs
A novel forensic approach for identifying abnormal LLM outputs
Precision Crash Detection for Traffic Safety
HybridMamba: Advanced temporal modeling for pinpointing crash moments
Revolutionizing Log Parsing with LLMs
Leveraging AI to transform unstructured logs into actionable security insights
Zero-Shot Error Detection for Data Integrity
Harnessing LLMs to Detect Errors in Tabular Data Without Training
Detecting LLM 'Quirks' That Supervision Misses
Using model internals to identify unexpected anomalies in AI systems
Smarter Anomaly Detection with MADLLM
Leveraging Language Models for Multivariate Time Series Analysis
Automating Root Cause Analysis with AI
Using LLMs to speed up incident resolution in cloud systems
AnomalyGen: Revolutionizing Log-Based Security
Using LLMs to Generate Realistic Anomaly Data for Better Detection
