With spaCy for entity extraction, Keras for intent classification, and more!

Over the past month, I wanted to look for a project that encompasses the entire data science end-to-end workflow — from the data pipeline, to deep learning, to deployment. It had to be challenging, but not pointlessly so — it still had to be something useful. It took a little ideation and divergent thinking, but when the idea of making a personal assistant came up, it didn’t take long for me to settle on it. Conversational assitants are everywhere. Even my university is currently using Dr. Chatbot to track the health status of its members as an effective way to monitor this current pandemic. And it just makes sense: chatbots are faster, easier to interact with, and is super useful especially for things that we just want a fast response on. In this day and age, being able to talk to a bot for help is starting to become the new standard. I personally believe bots are the future because they just make our lives so much easier. Chatbots are also a key component in Robotic Process Automation.

Now I want to introduceEVE bot, my robot designed to Enhance Virtual Engagement (see what I did there) for the Apple Support team on Twitter. Although this methodology is used to support Apple products, it honestly could be applied to any domain you can think of where a chatbot would be useful.

Here’s my demo video for EVE. (And here’s my Github repo for this project)

#nlp #named-entity-recognition #deep-learning #chatbots #customer-service #complete guide to building a chatbot with deep learning