1

I am trying to use UNSW-NB15 to train a model. After the model is trained, I would like to use the model on live network data. I began creating this using a supervised LSTM but started wondering about handling the data from the network and the necessity to create a data pipeline that preprocesses network data to get it in a manner similar to the UNSW-nb15 dataset. This seemed impractical to me as this would most likely mean going through data manually with each network data source. I am thinking that an unsupervised model may be better for my purposes. I still wanted to use LSTM but I'm finding very little in terms of information for creating an unsupervised lstm model in keras. Read a paper suggesting using BINGO (Binary Information gain optimization) or NEO (nonparametric entropy optimization) to train the lstm model. I am not certain how this can be done in keras. I am unable to find such functions there. (I will search python libraries though). Any suggestions?

I am still researching.

Alvinus Melius
  • 11
  • 1

1 Answers1

0

Sounds as if you're looking at "Unsupervised Learning in Recurrent Neural Networks" (Klapper-Rybicka et al 2001). I'm just a statistician so my ML expertise is weak, but still, it seems to me that's pretty old in ML terms; I doubt you'll find modern implementations of these exact ideas. Furthermore, on a quick read, it seems that the ideas behind NEO in particular are basically a proto-VAE — that is, superseded by modern models.

You'd do better IMO to look at https://arxiv.org/abs/1511.06349 (implemented in https://github.com/twairball/keras_lstm_vae) or https://arxiv.org/abs/1502.04681 (implemented at https://github.com/mansimov/unsupervised-videos). Neither of these optimizes for binary discrimination as BINGO does; but nevertheless, they should give a good starting point.

(I'm writing this answer because I just started from the same point you're at, and I want to document my findings for the next person who ends up here.)

Jameson Quinn
  • 1,060
  • 15
  • 21