For this activity, I was able to further deepen my understanding on Natural Language Processing. I learnt that textual data is not just fed to a model raw, but it is fed into some kind of process that converts it from words, that we humans can easily grasp and understand, to a representation that computers can easily understand and analyze. There are even some methodologies that I have encountered from my previous Elective course, like one-hot encoding, that could be applied to textual data to get a representation. Furthermore, it was satisfying how I was able to implement a word embeddings representation through my own search of Python machine learning libraries. It was fascinating that numbers generated by these processes can somehow, not only signify a word, but also context in the case of word embeddings.
Other than that, the activity once again reinforced the fact that steps and techniques done in a machine learning process have proper thought and justification put into it in a case-by-case basis. If I were to implement my own NLP process pipeline, I would have to make sure that I have picked the right text representation that suits the needs of my project.