Implement Gaussian Mixture Model using keras

Question

I am trying to implement Gaussian Mixture Model using keras with tensorflow backend. Is there any guide or example on how to implement it?

Do you want your network output to be a mixture of Gaussian or simply train GMM using Keras? — Marcin Możejko, Feb 27 '17 at 09:18
@MarcinMożejko , simply train GMM using keras. In similarly way as its done in [here](https://github.com/tensorflow/tensorflow/blob/master/tensorflow/contrib/factorization/python/ops/gmm.py) — Rahul, Feb 27 '17 at 09:56

score 3 · Answer 1 · answered Feb 27 '17 at 08:40

3

Are you sure that it is what you want? you want to integrate a GMM into a neural network?

Tensorflow and Keras are libraries to create, train and use neural networks models. The Gaussian Mixture Model is not a neural network.

answered Feb 27 '17 at 08:40

Nassim Ben

Yes I know that Gaussian MIxture Model is not neural network. But you can find a GMM in tensorflow [here](https://github.com/tensorflow/tensorflow/blob/master/tensorflow/contrib/factorization/python/ops/gmm.py). In a similar fashion, I want to implement GMM in keras – Rahul Feb 27 '17 at 09:58
2

@Rahul No, TF is a genera library for symbolic computation, while Keras is a specific library for neural networks. You can't implement a GMM on Keras while you can on top of TensorFlow or Theano. – Dr. Snoopy Feb 27 '17 at 11:20
@Rahul It's possible - but it would be extremely cumbersome. Why do you want to implement GMM in Keras? – Marcin Możejko Mar 03 '17 at 13:09
I don't know why the original poster wanted a GMM in Keras, but there are cases in which one wants a hybrid network where a part of the input is encoded as a GMM. – JWLM Feb 07 '19 at 15:05
Why is it so hard to believe that someone would want a GMM on graph? I want one too, the parameters are differentiable, there is no reason it can't be done and you don't need to know why. Classic stackoverflow. Thanks all for nitpicking instead of answering the question. – markemus Sep 28 '22 at 08:59

1 Answers1