D

Domurls BERT

Developed by amahdaouy
DomURLs_BERT is a BERT-based pretrained encoder specifically designed for detecting and classifying suspicious/malicious domains and URLs.
Downloads 725
Release Time : 8/5/2024

Model Overview

The model is pretrained on a large-scale corpus containing multilingual URLs, domains, and Domain Generation Algorithms (DGA) through masked language modeling objectives, making it suitable for malicious domain and URL detection and classification tasks in the cybersecurity field.

Model Features

Multilingual Support
The model is pretrained on a corpus containing multilingual URLs and domains, enabling it to process domains and URLs in different languages.
Malicious Domain Detection
Detection capabilities specifically designed for cybersecurity threats such as phishing attacks, malware, DGA, and DNS tunneling.
BERT-based Architecture
Leverages BERT's powerful contextual understanding to effectively capture potential threat features in domains and URLs.

Model Capabilities

Malicious Domain Detection
URL Classification
Phishing Attack Identification
DGA Domain Detection
DNS Tunneling Detection

Use Cases

Cybersecurity
Phishing Website Detection
Identify phishing URLs disguised as legitimate websites
Outperforms existing character-based deep learning models in evaluations
Malware Domain Detection
Detect domains used for spreading malware
Performs better than cybersecurity-specific BERT models on multiple tasks
Threat Intelligence
DGA Domain Identification
Detect malicious domains generated by Domain Generation Algorithms
Featured Recommended AI Models
AIbase
Empowering the Future, Your AI Solution Knowledge Base
Š 2025AIbase