What’s an Embedding?

Embedding (also called Vector Embeddings) is a series of vectors providing a mathematical representation of words or sentences. The vectors capture the semantic meaning and context of the words or phrases.

The advantage of Vector Embeddings is that they permit us to match and analyse words and phrases mathematically, enabling us to perform tasks equivalent to natural language processing, text classification, and data retrieval. Additionally they allow us to discover similarities and relationships between words, even in the event that they are quite different.

For instance, the Vector Embeddings for “dog” and “cat” can be much closer to every apart from that of “dog” and “banana”. Despite the fact that dog and cat don’t share the precise spelling, meaning, and space within the dictionary, they aren’t synonyms. Nonetheless, they share characteristics like they’re each four-legged, can each be kept as pets, could be trained etc.

Now that we understand the fundamental concept of Embeddings let’s have a look at how we are able to use this to simplify painful Data Management activities.