Summary
Here we discussed some of the mathematical background as well as some implementations we did not cover in the other sections. First we discussed the mathematical notation for scalars, vectors, matrices and tensors. Then we discussed various operations performed on these data structures, such as, matrix multiplication and inversion. Next, we discussed various terminology that is useful for understanding probabilistic machine learning such as, probability density functions, joint probability, marginal probability and Bayes rule. Afterwards, we moved our discussion to cover various implementations that we did not visit in the other chapters. We learnt how to use Keras; a high-level TensorFlow library to implement a CNN. Then we discussed how we can efficiently implement a neural machine translator with the seq2seq library in TensorFlow, compared to the implementation we discussed in Chapter 10, Sequence-to-Sequence Learning – Neural Machine Translation...