We're going to continue looking at Keras, but also at Recurrent Neural Nets and maybe LSTMs. It's about time.
Of course, these aren't trivial to jump into! And there's no Keras examples of Simple RNNs, so we try to shim in a Theano example. You decide how well it went...