A dataset to facilitate automated workflow analysis.

Area of Science:

• Computer Science
• Cybersecurity
• Machine Learning

Background:

• Manually annotating data for ground truth is time-consuming and expensive, limiting dataset availability for complex use cases.
• A lack of ground truth hinders automated situation awareness and understanding of cybersecurity breach consequences.
• Existing datasets are often niche, making cross-disciplinary applications difficult.

Purpose of the Study:

• To establish a ground truth for an email dataset to support automated workflow analysis.
• To bridge the gap between semantic mission goals and observable cyberspace events.
• To enable reliable, automated situation awareness in dynamic cybersecurity scenarios.

Main Methods:

• Developed a ground truth by assigning semantic labels to emails within an open-source software project dataset.
• Defined sequential email arrangements to represent observed actions and tasks.
• Utilized identified sequences to create template workflows and business process models.

Main Results:

• Successfully established a ground truth for a novel email dataset.
• Demonstrated the use of this dataset for workflow discovery.
• Showcased the potential for improved system accuracy in automated analysis.

Conclusions:

• Automated workflow analysis, supported by curated ground truth datasets, is crucial for cybersecurity.
• The developed dataset and methodology facilitate the understanding of persona-driven tasks in cyberspace.
• This work addresses a key barrier in achieving automated situation awareness and mitigating cybersecurity risks.