DNS-GT: A Graph-based Transformer Approach to Learn Embeddings of Domain Names from DNS Queries

Dimensions overall score 8.0

GitHub Code Pulse

No public code linked for this paper yet.

• Current embedding-based methods for DNS traffic do not properly consider contextual information among DNS queries.limitation 85%
• The code for DNS-GT is publicly available.technical 95%
• DNS-GT is a novel Transformer-based model designed to learn embeddings for domain names from sequences of DNS queries.method 95%
• The DNS-GT model is pre-trained in a self-supervised fashion to learn the general behavior of DNS activity.method 95%
• DNS-GT can be finetuned on specific downstream tasks, leveraging interactions with other relevant queries in a sequence.method 95%
• The proposed DNS-GT method achieves better results compared to relevant baselines on domain name classification tasks.result 90%
• The proposed DNS-GT method achieves better results compared to relevant baselines on botnet detection tasks.result 90%

Competitive landscape

Competitor map is still being generated for this paper. Enable generation or check back soon.

Startup potential card