Named-Entity-Recognition for Cybersecurity

• This project is about building a NLP model that can be used for Named-Entity-Recognition in the field of Cybersecurity.
• distilbert-base-uncased pretrained model was used in this project. This model is fine-tuned for this project's purpose.
• To open the project, Click on the above google colab badge.

Dataset

• MITRE Dataset is taken to train the model. The dataset is available in this repository as MITRE.zip. or Click here to download the dataset.
• MITRE dataset is also uploaded to huggingface. click here to go there.
• You can also import the dataset by running the below code.

from datasets import load_dataset

dataset = load_dataset("bnsapa/cybersecurity-ner")

Metrics

Huggingface

• The finetuned model is available on the huggingface. click here to go there.

Example

I typed abcde is a computer malware. I defined the context such that it implies abcde is virus and model is able to capture that.