# SecureReg: Combining NLP and MLP for Enhanced Detection of Malicious Domain Name Registrations

arXiv: https://arxiv.org/abs/2401.03196 · HTML: https://mlsanity.com/paper/2401.03196/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Builds on 8 of its 32 references that are in the map.
- First posted on arXiv 6 Jan 2024.

## Builds on (most cited first, 8 of 8)

- [BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding](https://mlsanity.com/paper/1810.04805/) (2018)
- [ALBERT: A Lite BERT for Self-supervised Learning of Language Representations](https://mlsanity.com/paper/1909.11942/) (2019)
- [MobileBERT: a Compact Task-Agnostic BERT for Resource-Limited Devices](https://mlsanity.com/paper/2004.02984/) (2020)
- [Recent Advances in Natural Language Processing via Large Pre-Trained Language Models: A Survey](https://mlsanity.com/paper/2111.01243/) (2021)
- [CANINE: Pre-training an Efficient Tokenization-Free Encoder for Language Representation](https://mlsanity.com/paper/2103.06874/) (2021)
- [Generating Fake Cyber Threat Intelligence Using Transformer-Based Models](https://mlsanity.com/paper/2102.04351/) (2021)
- [MalBERT: Using Transformers for Cybersecurity and Malicious Software Detection](https://mlsanity.com/paper/2103.03806/) (2021)
- [Less is More: Robust and Novel Features for Malicious Domain Detection](https://mlsanity.com/paper/2006.01449/) (2020)

## Built on by (most cited first, 1 of 1)

- [Real-PGDN: A Two-level Classification Method for Full-Process Recognition of Newly Registered Pornographic and Gambling Domain Names](https://mlsanity.com/paper/2511.22215/) (2025)

## Abstract (arXiv)

The escalating landscape of cyber threats, characterized by the registration of thousands of new domains daily for large-scale Internet attacks such as spam, phishing, and drive-by downloads, underscores the imperative for innovative detection methodologies. This paper introduces a cutting-edge approach for identifying suspicious domains at the onset of the registration process. The accompanying data pipeline generates crucial features by comparing new domains to registered domains, emphasizing the crucial similarity score. The proposed system analyzes semantic and numerical attributes by leveraging a novel combination of Natural Language Processing (NLP) techniques, including a pretrained CANINE model and Multilayer Perceptron (MLP) models, providing a robust solution for early threat detection. This integrated Pretrained NLP (CANINE) + MLP model showcases the outstanding performance, surpassing both individual pretrained NLP models and standalone MLP models. With an F1 score of 84.86% and an accuracy of 84.95% on the SecureReg dataset, it effectively detects malicious domain registrations. The findings demonstrate the effectiveness of the integrated approach and contribute to the ongoing efforts to develop proactive strategies to mitigate the risks associated with illicit online activities through the early identification of suspicious domain registrations.