My learning from pitting the Beast against the Grand Old Man!

The first and foremost challenge for data scientists in NLP tasks, after data cleaning, is to figure out how to represent their text mathematically. I was faced with this same challenge while working on building an open-domain semantic search based FAQ bot.

Embedding Methods

A Practical Consideration

The Verdict

…

#tfidf-vectorizer #similarity-search #data-science #bert #chatbots #bert vs tf-idf embeddings in an enterprise chatbot