vector embedding
noun

Definition:
A numerical representation of data—such as words, sentences, or entire documents—mapped into a multi-dimensional space where semantic relationships and contextual similarities can be mathematically quantified. Vector embeddings are used by machine learning models, especially large language models (LLMs), to understand and compare the meanings of different inputs based on their proximity within the vector space.

Usage:
“By analyzing vector embeddings, the model determined that ‘doctor’ and ‘physician’ were contextually similar.”

Compare:
Word2Vec, Sentence Embeddings, Semantic Vectors, Latent Space Representation

First Known Use:
Early 2010s, popularized through neural network-based natural language processing techniques such as Word2Vec and GloVe.