With advancements in machine learning (ML), deep learning (DL) and natural language processing (NLP), traditional conversational chatbots have now become advanced and are called contextual chatbots.
A conetextual chatbot understands the context of the conversation by obseving the pattern of communication between itself and the user .It not only keep track of the current state of the conversation,but also what has been said before.The advaned chatbots remember what a user asked or said previously,and deliver amore relevant answer according to that information.
Instead of relying on a bunch of if/else statements, a contextual chatbot uses an ML model trained on example conversations to decide what to do next.
If you have explored virtual assistants such as Apple’s Siri, Amazon’s Alexa, etc., you would have noticed that they are much more than a ‘speech-based search engine’. Apart from searching for information (e.g., from Wikipedia, YouTube, Google, etc.), they can ‘talk’ to you in natural language.
Conversation, or dialogue is a fundamental aspect of human language, and arguably the most interseting challenge in building truly inteligent NLP systems.
In this module ,you’ll learn to build a domain- specific chatbot using anopen-source framework for building conversational bots`, namely, Rasa. Specifically, in this session, you’ll learn to build a restaurant search bot, although the skills that you will learn can be used to build almost any domain-specific bot.
In the following video, Aiana will explain the reason why we are using the Rasa framework over alternative frameworks such as IBM Watson, Wit.ai, Amazon Lex, etc., for building our chatbot.
To summarise,any conversational system has primarily two components:
- Natural Language Understanding ,or NLU
- Dialogue Management System
which carries out the overall conversation
In Rasa, these two components are named Rasa NLU and Rasa Core, respectively.