What is DarkBERT?

A large language model to understand the Dark Web

Researchers at the Korea Advanced Institute of Science & Technology have pre-trained a large language model on documents acquired from the Dark Web to help fight global cybercrime.

Here are some key points:

• DarkBERT is a large language model (LLM) designed to handle the Dark Web's enormous lexical and structural variation.

• In terms of ransomware leak detection, notable thread detection, and threat phrase inference, the researchers believe it beats other pre-trained language models.

• The workload of cybersecurity specialists could be reduced by automating such analyses.