{"id":4450,"date":"2023-12-13T08:00:00","date_gmt":"2023-12-13T14:00:00","guid":{"rendered":"https:\/\/baylor.ai\/?p=4450"},"modified":"2024-07-19T18:05:59","modified_gmt":"2024-07-19T23:05:59","slug":"creation-and-analysis-of-an-nlu-dataset-for-dod-cybersecurity-policies","status":"publish","type":"post","link":"https:\/\/lab.rivas.ai\/?p=4450","title":{"rendered":"Creation and Analysis of an NLU Dataset for DoD Cybersecurity Policies"},"content":{"rendered":"\n
\"\"<\/figure>\n\n\n\n

Comprehending and implementing robust policies is crucial in cybersecurity. In our lab, Ernesto Quevedo et al. recently released a paper, Creation and Analysis of a Natural Language Understanding Dataset for DoD Cybersecurity Policies (CSIAC-DoDIN V1.0)<\/em><\/a>, which introduces a groundbreaking dataset to aid in this endeavor. This dataset bridges a significant gap in Legal Natural Language Processing (NLP) by providing structured data specifically focused on cybersecurity policies.<\/p>\n\n\n\n

Dataset Overview<\/h4>\n\n\n\n

The CSIAC-DoDIN V1.0 dataset encompasses a wide array of cybersecurity-related policies, responsibilities, and procedures from the Department of Defense (DoD). Unlike existing datasets that focus primarily on privacy policies, this dataset includes detailed guidelines, strategies, and procedures essential for cybersecurity.<\/p>\n\n\n\n

Key Contributions<\/h4>\n\n\n\n
    \n
  1. Novel Dataset<\/strong>: This dataset is the first to include comprehensive cybersecurity policies, guidelines, and procedures.<\/li>\n\n\n\n
  2. Baseline Models<\/strong>: The paper provides baseline performance metrics using transformer-based models such as BERT, RoBERTa, Legal-BERT, and PrivBERT.<\/li>\n\n\n\n
  3. Open Access<\/strong>: The dataset and code are publicly available<\/a>, encouraging further research and collaboration.<\/li>\n<\/ol>\n\n\n\n

    Experiments and Results<\/h4>\n\n\n\n

    Our team of researchers evaluated several transformer-based models on this dataset:<\/p>\n\n\n\n