Which statement best describes the purpose of word embeddings?

Which statement best describes the purpose of word embeddings?

• A. Represent tokens as vectors capturing syntactic and semantic information
• B. Represent documents as counts
• C. Represent characters as scalars
• D. Increase model training time

Answer: (A)

Word embeddings convert words into continuous vectors that encode how words are used in language, capturing both syntactic and semantic relationships. This means similar words end up with similar vectors, and relationships between words can be reflected by simple vector operations (for example, analogies like king minus man plus woman being close to queen). This representation enables models to measure word similarity, compare meanings, and learn from word context during training. Other options miss the core purpose: representing documents as counts describes a bag-of-words approach that loses word-level nuance and order; representing characters as scalars isn’t how embeddings work (we map tokens to vectors, not single numbers); and simply increasing training time isn’t a purpose of embeddings—it’s a consequence or side effect, not what the embeddings are for.

Word embeddings convert words into continuous vectors that encode how words are used in language, capturing both syntactic and semantic relationships. This means similar words end up with similar vectors, and relationships between words can be reflected by simple vector operations (for example, analogies like king minus man plus woman being close to queen). This representation enables models to measure word similarity, compare meanings, and learn from word context during training.

Other options miss the core purpose: representing documents as counts describes a bag-of-words approach that loses word-level nuance and order; representing characters as scalars isn’t how embeddings work (we map tokens to vectors, not single numbers); and simply increasing training time isn’t a purpose of embeddings—it’s a consequence or side effect, not what the embeddings are for.