A Comprehensive Dataset Towards Hands-on Experience Enhancement in a Research-Involved Cybersecurity Program

Undergraduate research activities have demonstrated their ability to enhance learning outcomes, improve student retention, and foster critical thinking. In particular, cybersecurity is a highly practical discipline, and relying solely on passive learning or observation is far from sufficient. By actively engaging in research, students can better comprehend and apply the knowledge that they have acquired, thus cultivating their ability to solve practical problems. On the other hand, we observe that well-established datasets hold paramount importance for AI-driven cybersecurity applications, while pertinent data are usually a precious and scarce resource. In this paper, we introduce a comprehensive dataset SecAtlas towards research-involved cybersecurity programs using real-world cases. The dataset construction provides students with various realistic samples of cyber-attacks and vulnerabilities, which can make them experience the complexity of cybersecurity and acquire relevant skills. More importantly, SecAtlas greatly facilitates the engagement of students in subsequent data-centric cybersecurity research. Finally, we conduct a set of studies to evaluate the collected dataset. The investigation results demonstrate the effectiveness of applying the proposed dataset to practical security projects.