numdouglas/cg-chatbotty

Implementation of a chatbot that learns from a script to classify queries and generate answers for interactions.

Context-Generative Chatbot

Simple, accurate and portable chatbot trained off a corpus of conversations to respond to real-time queries.

• Uses Tensorflow Deep Neural Network to model the script and have responses that will be available and vary based on context.
• For classification, regression analysis is used, backed by the modern RELU which minimizes backpropagation of errors.
• NLP is used to create features which will be trained off Tensorflow's DNN's to attain a model with labels and context.
• The chatbot is made to be embedabble to any webpage as a chat bubble, running multiple instances from a single server, via a Flask, REST and Vue.js Framework.




Here's the chatbot enjoying a casual conversation and then going deep.