Introduction
One of the latest trends in deep NLP is the creation of conversational agents, also knowns as chatbots. A chatbot is a text-based dialogue system that understands human language and can hold a real conversation with people. Many companies use these systems to interact with its customers to obtain information and feedback, for example, opinions on a new product launch.
Chatbots are used as assistants, for example, Siri, Alexa, and Google Home. These can give us real-time information about the weather or traffic.
At this point, the question is how can bots understand us? In the previous chapters, we have reviewed language models and how they work. However, the most important thing in language models (LMs) is the position of a word in a sentence. Each word has a certain probability of appearing in a sentence, depending on the words already in that sentence. But the probability distribution approach is not a good fit for this task. In this case, we need to understand the meaning...
