Receive a weekly summary and discussion of the top papers of the week by leading researchers in the field.

In Computers & security

This paper examines the transition in the cyber-security discipline induced by the ongoing COVID-19 pandemic. Using the classical information retrieval techniques, a more than twenty thousand documents are analyzed for the cyber content. In particular, we build the topic models using the Latent Dirichlet Allocation (LDA) unsupervised machine learning algorithm. The literature corpus is build through a uniform keyword search process made on the scholarly and the non-scholarly platforms filtered through the years 2010-2021. To qualitatively know the impact of COVID-19 pandemic on cyber-security, and perform a trend analysis of key themes, we organize the entire corpus into various (combination of) categories based on time period and whether the literature has undergone peer review process. Based on the weighted distribution of keywords in the aggregated corpus, we identify the key themes. While in the pre-COVID-19 period, the topics of cyber-threats to technology, privacy policy, blockchain remain popular, in the post-COVID-19 period, focus has shifted to challenges directly or indirectly brought by the pandemic. In particular, we observe post-COVID-19 cyber-security themes of privacy in healthcare, cyber insurance, cyber risks in supply chain gaining recognition. Few cyber-topics such as of malware, control system security remain important in perpetuity. We believe our work represents the evolving nature of the cyber-security discipline and reaffirms the need to tailor appropriate interventions by noting the key trends.

Kumar Rajesh, Sharma Siddharth, Vachhani Chirag, Yadav Nitish


COVID-19 pandemic, Cyber-security trends, Latent Dirichlet Allocation, Topic modeling, Trend analysis, Unsupervised machine learning