VectorDatabase

What is semantic search?
when you search on google about apple, how it knows that this apple is either related to company or fruit.
Concept behind is gonna be semantic search.
It figures out user intent, embeddings are used there. Words are created into numerical representations.
Those numerical representations.

First technique is we have to think about features in that particular word like either it has calories, location, is it fruit, is 
it related to phones, 


First is word we search , 2nd is context

if you search about fruits like apple ->
related_to_phones 0
related_to_fruits 1
revenue 0 (because it is related to fruit)


Cosine function technique is also that matches vector result if it is neer to one then added to result


BERT,GPT are techniques to gather numerical representation of specific word

We use hashing to get those similar vectors


LOCALITY SENSITIVE HASHING