Detecting phishing with Recurrent Neural Networks

Phishing is a simple kind of cyberattack, usually in the form of an email, that has been more frequent than ever. Its goal is to deceive people, by various means, into giving personal sensitive information such as passwords.

Security research has been extensive in this matter and recently neural networks have been implemented to stop the damaging effects of these kinds of attacks.

In this talk we'll discuss the efficacy of Recurrent Neural Networks with LSTM architecture on phishing detection

João Pereira iis a Deliverability and Security Engineer at E-goi, a multi-channel marketing automation platform. João is finishing his degree in Informatics Engineering in ISEP. He works with Data Science since 2018 and during an internship at E-goi created a phishing detection system based on Deep Learning techniques.

Transcript

1. Multichannel Marketing Platform Phishing Detection with 
 Recurrent Neural Networks João Pereira
2. The world E-goi +15 Years
 experience 500.000
 users 700.000.000
 SMS sent in 2018 100.000.000.000
 Emails sent in 2018
3. What was accomplished? F1-Score Recall Precision 98.0% 98.2% 98.4%
4. About Phishing
5. What do you look for? Sender Email doesn’t make sense Sense of urgency
 Call for immediate action Generic greeting Often poorly written ....
6. The Data issue “Data! Data! Data! I can't make bricks without clay.” - Sir Arthur Conan Doyle
7. The tools for the job Python Google Colab Keras
8. [ 4, 23, 12, 52, 65, 8, 24, 23, …, 14 ] Tokenize Making the data readable Generate Embeddings
9. How about neural networks? Feed Forward Network Recurrent Network
10. Understanding the Architecture Great explanation at: https://colah.github.io/posts/2015-08-Understanding-LSTMs/
11. Understanding the Architecture Great explanation at: https://colah.github.io/posts/2015-08-Understanding-LSTMs/
12. Forget Gate
13. Forget Gate
14. Input Gate
15. Input Gate
16. Cell State
17. Cell State
18. Output Gate
19. Output Gate
20. Thank you!