Cameron McGinley
4 Comments
Leave a Reply
You must be logged in to post a comment.
Kansas Undergraduate Research Day
Online March 2, 3 & 4, 2021
Kansas Undergraduate Research Day
Online March 2, 3 & 4, 2021
You must be logged in to post a comment.
Wow! This is very interesting. What motivated you to study this?
Thanks for the question! I was motivated to study this because deep learning has rapidly advanced over the past few years, and it excels in detection tasks. Attackers are constantly changing their techniques to circumvent detection, but deep learning is incredibly effective at adapting to any changes while reaching very high accuracy, which would help stop the growth in phishing attacks.
Very interesting topic. What are your next steps with this research?
Thanks! Dr. Salinas and I have continued research on this topic already, and we have focused on using deep learning techniques to automatically generate responses to malicious emails in order to deceive the attacker. For this, we are using a finetuned version of OpenAI’s GPT-2. It was important to research detection methods first, because if we automatically respond to emails we had better be sure we are responding to the right ones.