Progress Report (Meeting Notes Here)

[JoeyTaubert]

Major changes to project path:

1. WebApp Format (Scrap idea of native GUI)

2. Instead of Network TAP/SPAN, possibly integrate with SIEM products to evade cost assocaited with TAP/SPAN. Putting this on back burner for now.

3. File storage, flat files (one line = complete data)

4. Proof of concept - Something simple: Given echo request, able to predict echo reply? 

Progress:

Developed a Recurrent Neural Network (RNN) (there are other names for this but I will be referring to it as a RNN.

• This is based on a YouTube video I found that created a RNN to predict the outcome of a basic XOR truth table.
• Planning on editing number of inputs/outputs to see if we can provide packet information and return packet information. Or something similar.

NOTE: This is just a stepping stone for my second model, a Long-Short Term Memory model (LSTM). The major improvement it will provide is that the network will be able to learn effectively given large datasets (see exploding/vanishing gradient problem). 

[JoeyTaubert]

Axum - WebApp Framework Proceeding with raw packets as input, but log analysis could be a fun fork of this File storage = flat files (for now, maybe NoSQL/MongoDB) PoC = In Progress