How to Train your Chatbot
Chatbots are key technologies that allows a business to respond to customer's query fast and efficiently. Many organizations are using chatbot to manage customer service response efficiently and requiring not too many human resource as support staffs. Mostly Chatbots has been used to answer basic queries.
But companies are generating a large amount of unstructured and structured text data. Also new images and videos are being processed. A lot of research has already gone into how to train chatbots on structured question answering text.
In this research we try to answer if chatbot can learn from these unstructured data and generate answer to questions in an unsupervised manner. Chatbot can also learn using OCR/image processing from images.
We propose a efficient framework to train chatbots and organization can do it on their own.
Outline/Structure of the Talk
What are Chatbots, Information retrieval, information extraction
Building a simple chatbot: framework - Supervised/Unsupervised
Augmenting chatbot with human customer support system - How to build the framework
Information extraction and knowledge representation
How to use knowledge representation for Question Answering
Online and continuous Learning
Organization will learn how to build and train their chatbots
How to maintain and integrate with customer support
Challenges, latest ML libraries and NLP techniques
The future of chatbots and Beyond
Useful for all
Prerequisites for Attendees