Vector embeddings are one of the most fascinating and useful concepts in machine learning. They are central to many NLP, recommendation, and search algorithms. If you’ve ever used things like recommendation engines, voice assistants, language translators, you’ve come across systems that rely on embeddings.
ML algorithms, like most software algorithms, need numbers to work with. Sometimes we have a dataset with columns of numeric values or values that can be translated into them (ordinal, categorical, etc). Other times we come across something more abstract like an entire document of text. We create vector embeddings, which are just lists of numbers, for data like this to perform various operations with them. A whole paragraph of text or any other object can be reduced to a vector. Even numerical data can be turned into vectors for easier operations.
This is a companion discussion topic for the original entry at https://www.pinecone.io/learn/vector-embeddings/