Application of Reinforcement Learning Yields More Robust Data Breach Controls for Working from Home

In the last 5 years, the United States has lost $13.2 billion to cyberattacks. Microsoft, FireEye, the Department of Defense, the Department of Homeland Security, and many other reputable organizations were victims of cyberattacks in 2020. Despite the advanced countermeasures used on office networks to prevent data breaches, attackers always find ways to steal data. With work from home as the new status quo and the lack of similar countermeasures on home networks, the resistance to cyberattacks is exceptionally inadequate. This research primarily focused on reducing data breaches resulting from telework. It created predictive models using a reinforcement learning methodology to improve the robustness of data breach controls for employees working from home. It leveraged data sources including domain risk factors (e.g., popularity, suspicions, citations), employee risk factors (e.g., background, behavior, associations), and content risk factors (e.g., personal identifiable information, sentiment, malware signatures). This research makes three significant contributions. First, it proved that a reinforcement learning model can significantly increase the accuracy of predicting data breaches. Second, it identified employee risk as a novel factor in preventing data breaches when threat intelligence is missing. Third, the predictive model of this research can integrate with home wireless routers, making it accessible to families, students, and small businesses that cannot afford expensive cybersecurity services. Thus, this research will help reduce data breaches for employees working from home.

Author

• Mallarapu, Ravinder

Language

• en

Keyword

• Artificial Intelligence
• Machine Learning
• Data Breaches
• Working from home
• Remote work
• Cybersecurity

Date created

• 2021

Type of Work

• Dissertation

License

• All rights reserved

Rights statement

• In Copyright

GW Unit

• Engineering Management

Degree

• D.Engr.

Advisor

• Sarkani, Shahryar
• Fossaceca, John

Committee Member(s)

• Fossaceca, John
• Etemadi, Amir
• Sarkani, Shahryar

Persistent URL

•  https://scholarspace.library.gwu.edu/etd/8k71nh81w