Build Your Own Chatbots That Know If You Are Angry




Build your own chatbots that know if you are angry

4 July 2019


Added 01-Jan-1970

Do you want to build and train your own chatbot?

In this workshop, through hands on coding exercises, participants will learn how to use an open source chatbot framework, RASA, in conjunction with an NLP library, NLTK, to build a chatbot that gathers customer feedback and generates a report regarding the sentiment of the text.

In the first part, we will introduce RASA, including how to build and train a chatbot using RASA. We will start from scratch, covering the necessary installation steps, then setting up the project, and building a simple working chatbot. These steps will include setting up a pipeline for the RASA NLU, learning how to write a training file for the NLU intent, using the custom form action to get and handle users’ feedback, and writing the stories that make up the dialogue flow.

In the second part (Optional, depending on time), we will cover the sentiment analysis tool NLTK. This will enable a deeper use of RASA, including more custom settings in the architecture. We will explain how sentiment analysis in NLTK works and integrate that sentiment analysis in the chatbot so that it can handle users differently according to their expressed sentiment. We will create a custom action that handles the sentiment analysis and stores the result, a measure of the positivity or negativity of the user dialogue that can be used to generate reports, or prompt different responses, etc.

We will work in groups (pair programming), although this workshop is for people who are confident programming in Python and who have a basic understanding of NLP. We encourage participants with different level of experience working in a group together.

Arrival and food 6:00pm
Separate into groups and presentation of the sponsor at 6:30pm
Workshop start at 6:45pm
Finish by 9:00pm