Of course Python, It's comes with large amount of inbuilt libraries. In these libraries some are for Artificial Intelligence and Deep Learning Such as Tensorflow (for neural network library), scikit-learn (for data analysis ), pylearn2 (more flexible than scikit-learn), etc. The list keeps going and never ends.
Thinking of programming issues, Matlab would provide a bit easier way to program your project as there are plenty of dedicated functions in it (more than 15000 functions in a recent version of Matlab, in several toolboxes). You may find out several functions in NN toolbox for your application.
However, both languages are high-level, and either can be considered for a researcher or developer. Loosely speaking, Matlab is widely known in academia, say in research areas, and Python is mostly known in the implementation stage, say in the developing areas.
When it comes to programming it depends with your knowledge in a particular language that you may apply. Matlab is rich with many functions and is sort of a self learning language unlike python. But still both can work well in communication Engineering
I would recommend Python and Keras (http://keras.io/).
Keras is an open source high-level neural network library written in Python and capable of running on top of Google's TensorFlow, Microsoft's CNTK, or Theano. In 2017, Google's TensorFlow team decided to support Keras in TensorFlow's core library. It's very user friendly, supports different types of networks and runs on both CPU and GPU.