New Transformer-based IP Geolocation Model Leverages LLMs for Enhanced Accuracy

Researchers have developed IPGeoAI, a novel deep learning model that uses Transformer architecture for more accurate city-level IP geolocation. This model reframes geolocation as a sequential modeling task, capturing hierarchical dependencies in IP subnet structures. It integrates unstructured semantic context from large language models (LLMs) to resolve geographic ambiguities, transforming raw Autonomous System descriptions into structured metadata. Extensive evaluations show IPGeoAI significantly outperforms leading vendors, achieving a 6% improvement in city-level accuracy and extending coverage to 100% of traffic. AI

IMPACT Enhances accuracy in IP geolocation services, potentially improving content delivery and digital rights enforcement.

RANK_REASON The cluster describes a novel deep learning model architecture presented in an arXiv paper. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.AI →

• Avinash Kadimisetty
• IPGeoAI
• IPv6
• large-language models
• Multi-Head Cross-Attention
• Transformer++
• Zero-Shot LLM Feature Extraction

• paper
• model release
• infra

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →